Paul Davies

______________________________________________________

Reply by Dieter Zeh:

>In an exchange with Jonathan Oppenheim, Dieter Zeh posed a question about
>applying quantum mechanics backward in time:
>
>'In particular, I do not understand what postselection (required for Bill
>Unruh's symmetry arguments) means if there is no ensemble to select from.'
>
>to which Oppenheim replied:
>
>' I think Bill's argument is just that quantum mechanics itself says nothing
>about a time direction, so in and of itself, it can't yield an argument
>which says the past is different from the future. But what IS "quantum
>mechanics in and of itself"? The Schroedinger equation and formal transition
>amplitudes are certainly insufficient to describe measurements.'

The last two sentences were mine (hdz) -- not Jonathan's. [MY APOLOGIES. I TRIED TO SIMPLIFY THE DIALOGUE, AND WAS CONFUSED BY THE PROLIFERATION OF > SYMBOLS. AS EDITOR OF THIS DISCUSSION, I UNDERSTOOD IT TO BE MY JOB TO TRY AND STREAMLINE THINGS.] Of course, we all agree that quantum dynamics is time-symmetric (including transition probabilities). This cannot be all he meant! I am also confused about who wrote the following paragraph. Sounds like a self-participatory Bill Unruh:

>Well, Bill Unruh now explains for himself what his argument is. I want to
>start by presenting his two key paragraphs, which is a succinct and eloquent
>encapsulation of his central position:

[I WROTE THIS PARAGRAPH IN AN ATTEMPT TO HELP READERS TO NAVIGATE THROUGH THIS COMPLICATED MULTIPLE EXCHANGE.] We will see ourselves what is eloquent or even a good argument. Since the next two paragraphs of the original comment were merely quoted from the subsequent "full text" (why?), I now leave them out and jump to this full text. [I DECIDED IT WAS USEFUL FOR READERS NOT WISHING TO FOLLOW THE MINUTAIE OF THE DISCUSSIONS TO HAVE THE KEY POSITION SPELLED OUT IN ADVANCE.]

>The full text of Unruh's remarks now follows:

So if this is not his own e-mail, when and where did he make them? [THE FULL TEXT THAT FOLLOWED IS EXACTLY WHAT BILL UNRUH SUBMITTED. I HAVE IN NO WAY EDITED HIS REMARKS, EXCEPT TO EXTRACT THE TWO PARAGRAPHS QUOTED AT THE BEGINNING, AND TO INTERJECT OCCASIONALLY.] At least I hope now to answer the genuine Bill Unruh!

>Measurements are not the question, nor were my comments an attempt to
>explain the "measurement problem". It is a question of setting
>conditions.

What is the difference? Do you "set" the conditions or do you just measure them? Can you set them without knowing (that is, having measured in some sense) what you have set? And would you have to change the conditions by a measurement (as is assumed even in the Copenhagen interpretation)? Instead of saying that you SET conditions before and AFTER you made the measurement, I would say that you made several measurements in a certain temporal order. Perhaps you are indeed thinking of measurements occurring "outside the laws of nature"? Nonetheless, there seems to be consensus that in quantum theory they have to CHANGE the physical state (in general). Whenever a measurement result (or a condition) is macroscopically documented, it BECOMES a "fact" (irreversibly). Measurements are said not to be complete if reversible. So time symmetry is gone as soon as you include the apparatus or the environment into your description because of the arising entanglement in spite of symmetric laws.

> In classical physics, conditions are called initial
>conditions, and can always be mapped back to being equivalent to initial
>conditions. But in general one wants to ask of the theory "Given certain
>known facts about the actual physical world, what can one say about
>other, currently unknown facts about the world?"

>The setting of the known is the setting of the conditions. Quantum
>mechanics says that the setting of conditions is not reducible to
>the setting of initial
>conditions.

Since quantum mechanics cannot speak, I assume it is Bill Unruh who says so.

>In particular, some conditions (eg conditions about the
>world after the occurence of the unknowns) produce results which depend
>on the temporal order in which they are imposed-- I gave an example in
>my first comment. Quantum mechanics is a set of rules allowing one to
>argue from those conditions to conclusions about the unknowns.
>In this view, the question is not "Well, how do you really know those
>things which you are setting as conditions" but rather "Given those
>conditions, what other conclusions can you draw."

Who "gives" conditions, and how?

>Although quantum mechanics is often presented in the guise of a
>classical field

No, a classical field is defined in space - not in a classical configuration space.

> theory (ie the Schredinger equation) leading one to the
>idea that what is important is the setting of initial conditions on that
>classical field ( the "amplitude"), this is an approach which has led to
>years of sterile arguments about the reality

Sterile as long as reality was assumed to be local. Now it seems to be very fertile (as you see from decoherence and many novel experiments). Entanglement has been observed in situations when it remains controllable. Why should it cease to exist when it becomes uncontrollable?

>of this field and the changes induced in this field by
>"measurements". On the other hand, regarded as a theory by which one
>extends known conditions (however arrived at) to other conclusions,
>quantum mechanics allows one to set
>conditions however one pleases, and in particular set them in the future
>as well as the past.

As a mathematician? In your example you had to use measurements.

>One of the ways in which one can do the calculation, which makes many
>people feel comfortable,

Aharonov and Vaidman? The latter expressed some sound ideas about the many minds interpretation.

> is to argue about there being some ensemble in
>the future from which one makes certain "post selections". But that is
>imposing a certain philosophical attitude on ones calculations.

I even think it is untenable.

> The calculations themselves give unambiguous answers (assuming the question
>and the conditions were unambiguous). There is no "selection of a wave
>function".

Whatever is selected, just tell us!

> There is simply a rule by which one proceeds from the
>conditions one sets (and the insights into the general dynamics of the
>system one is interested in), to the predictions (definite or
>probabilistic) about other outcomes which are, by hypothesis, unknown.
>That rule is embodied by, say, the two-wavefunction formalism of Aharonov,
>or the decoherence functional of Aharonov, Bergman and Liebowitz, or
>GellMann and Harle.

This must be another decoherence than the usual one in terms of entanglement (of the wave function) that has been experimentally confirmed.

>The wave function is not a thing which lives in the world.

You mean: not in space, but the "real" word has been shown to be nonlocal.

> It is a tool
>used by the theory to make those inferences from the known to the
>unknown. Once one knows more, the wave function changes, since it is
>only there to reflect within the theory the knowledge one assumes one
>has about the world.

This interpretation of the wave function as representing mere knowledge is untenable, as I describe in my chapter, for example.

>Unruh now challenges two other points made by Zeh:

Who is writing this, again? [PAUL DAVIES. I WAS TRYING TO BE HELPFUL TO THE READER.]

> > Several good textbooks discuss the irreversible nature of
> > measurements. Paul Davies' book of 1974/77 contains a chapter on Time
> > Asymmetry in Quantum Mechanics. In the third and fourth edition of my
> > book, the chapter on the Quantum Mechanical Arrow of Time contains a
> > section on The Time Arrow in Various Interpretations of Quantum
> > Theory (see www.time-direction.de), where I have discussed (and
> > criticized) Aharonov's concept of postselection.
>
>But this is irrelevant to the issue.

To what issue?

> Whether or not the theory actually
>contains within itself the possibility of describing how the world
>actualises itself, or contains a theory of how knowledge is obtained
>from the world, or needs a theory of conciousness to be complete, are
>further interesting metaquestions, which really have little bearing on how
>quantum mechanics itself works as a theory. Most of the Quantum Arrow of
>time is tautological.

Great claim, but what IS the quantum arrow if not the irreversibility of measurements?

> One sets initial conditions, and then is surprised
>that the past and the future differ.

Who is "one"?? The observed decoherence confirms the presumed initial condition that entanglement is absent (or negligible for the future), while future entanglement is not neglible for the past. It is required to describe the correct individual physical state.

> In classical physics, past and future conditions are equivalent,

Mathematically, but neither actually nor operationally.

> as any conditions are equivalent to initial
>conditions. This has made us believe that this should be true of physics in
>general-- that all conditions are initial. This is not true in quantum
>theory, and it is the imposition of such biases which I think lead to the
>belief that there is thus a difference between past and future buried in the
>theory itself. As Aharonov amongst others has made clear, there isn't.

I disagree. But in my opinion the difference is indeed not in the theory itself, but in this actual world (that can be described by its time-dependent wave function).

> > Freeman Dyson probably refers to the irreversible transition from
> > potentiality to actuality used in the Copenhagen interpretation. For
> > example, Pauli once declared that "the occurrence of a definite
> > electron position is a CREATION outside the laws of nature". In the
> > theory of decoherence, the click in the counter is instead described
> > by the "irreversibly" arising entanglement (an effect that has been
> > observed, as Serge Haroche will probably tell us), based on the
> > Schroedinger equation AND on a time-asymmetric environment. No
> > measurement can be regarded as complete if it remains reversible.
> > However, if the state is (irrerversibly) CHANGED in a measurement,
> > you cannot postSELECT it any more.
>
>Decoherence comes no closer to explaining the mystery of why the
>actualisation of the world manifests itself in definite outcomes
>while the theory only talks of probabilities, than does say
>Copenhagen.
> It may well explain why certain types of coherence, which the most
>naive forms of quantum mechanical analysis might predict (eg,
>experiments sensitive to the exact phase between live and dead cat
>states), are unobservable. It says nothing about why that cat is
>experienced to be either alive or dead when
>even the most sophisticated analysis predicts 50-50 probablity for the two
>possibilities.

I hope that Wojciech Zurek and others will soon join the party!

> (Everett "explains" it by denying the experience.)

This is a severe misunderstanding. Everett concludes precisely (by using the Schroedinger equation) that both situations "exist" in one superposition. Decoherence explains that only the world components with these two "pointer states" are dynamically autonomous, such that a dead or alive cat (but none of its superpositions) can be experienced by an appropriately defined local observer. Why is this conclusion trivial or naive after two generations of physicists overlooked the role of the (time-asymmetric) environment? Probabilities are only meaningful as frequencies in series of experiments.

>Furthermore, the state is not a physical item which is changed in
>a measurement.

I doubt it. That's what my chapter is about. Are you denying the individual physical meaning of superpositions, such as a K_long?

>Our knowledge might be changed, and as a reflection of
>that change, one might (depending on what conclusions one wished to
>derive) therefore also change that representation of our knowledge within
>the theory. But this is not the change of anything physical within the
>system under consideration.
>
>[PD interjects here. Bill, this last paragraph confuses me. Is it not the
>case that after a measurement the wavefucntion is generally
>different, and that this difference (after 'the collapse,' to use a
>non-PC description)

What is a non-PC description? (I am writing all descriptions on a Mac! ;-) ) Otherwise I agree, of course...

AH! SORRY FOR THE CONFUSION. 'PC' IS NOW WIDELY TAKEN TO STAND FOR 'POLITICALLY CORRECT.' PERHAPS MY TOUCH OF HUMOR IS INAPPROPRIATE...]

>distinctly affects the actual physical behavior of the system subsequently?
>To be sure, our knowledge jumps, but so do the particles.]

... except that I neither believe in (discrete) particles, nor in jumps.

>This sort of epistomological approach to quantum theory of course does
>not give one a very good clue as to how we should think of the world.
>The notion that there is really a wavefunction out there, which alters
>and changes as time goes by and as we make those mysterious things
>called measurements, is, naively, much more comforting and in keeping
>with mechanistic prejudices

There is nothing mechanistic in a nonlocal wave function. Do you mean deterministic? Even that would be irrelevant in practice for us "branching" quantum objects. You may as well assume an indeterministic modification of the Schroedinger equation (still describing reality) if you prefer. Rather, the prejudice is the assumption that reality must be described by local classical concepts.

> with how the world should operate.
>But the world is a "smoky dragon" in John's words, of which our
>experience only gives us glimpses here and there. To try to dispel the
>smoke by clothing the dragon with items like wave functions I think just
>hides the mystery of the world from us, and make us think that it is
>really not as strange as it is.

Well, but there is no reason to create even more "Schall und Rauch" as soon as part of the dragon may become recognizable -- even in the form of a universal wave function.

Dieter Zeh (so you know who wrote the non- > parts of this comment)

FINAL COMMENT. IF ANYONE WISHES TO JOIN THIS DICUSSION WOULD THEY PLEASE EXTRACT AND RETYPE THE APPROPRIATE SECTIONS FROM THE TEXT, AND LABEL CAREFULLY WHOSE WORDS THEY ARE. WE MUST AVOID CREATING STRINGS OF >>>>>>> THAT CONFUSE PEOPLE.]

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
This email list is part of the Science & Ultimate Reality Symposium in honor of John Archibald Wheeler, March 15-18, 2002 in Princeton, N.J. For more information go to: http://www.metanexus.net/ultimate_reality. This list is moderated by Paul Davies. Please feel free to forward these messages to friends, colleagues, and students.

To subscribe, unsubscribe, or access the online archives, please go to:http://listserv.metanexus.net/metanexus/archives/wheeler.html.

This publication is hosted by Metanexus Online http://www.metanexus.net. The views expressed here do not necessarily reflect those of Metanexus or its sponsors.

Metanexus welcomes submissions between 1000 to 3000 words of essays and book reviews that seek to explore and interpret science and religion in original and insightful ways for a general educated audience. Previous columns give a good indication of the topical range and tone for acceptable essays. Please send all inquiries and submissions to . Metanexus consists of a number of topically focused forums (Anthropos, Bios, Cogito, Cosmos, Salus, Sophia, and Techne) and periodic HTML enriched composite digests from each of the lists.

Copyright notice: Except when otherwise noted, articles may be forwarded, quoted, or republished in full with attribution to the author of the column and "Metanexus: The Online Forum on Religion and Science ". Republication for commercial purposes in print or electronic format requires the permission of the author. Copyright 1998, 1999, 2000, 2001, 2002, 2003, 2004 by Metanexus Institute.

Published   2002.02.21