Transactional interpretation
30 views
Skip to first unread message
Alan Forrester
Jan 20, 2013, 2:39:20 PM1/20/13
to beginning-of-infinity@googlegroups.com Infinity
Advocates of the transactional interpretation (TI) claim that their theory is better than the many world interpretation (MWI), see:
http://arxiv.org/abs/1001.2867
Let's have a look at the substance of their claims.
They describe the EPR experiment. They then say
> The more general consideration of dual or multiple measurements on entangled quantum systems, as in EPR experiments, has not been adequately addressed. The interpretational problem with such systems is that while each individual measurement may have many possible outcomes, because of conservation laws acting in entangled systems, only certain outcomes of a second measurement are permitted for any particular outcome of the first measurement. Only pairs of outcomes are permitted for which the conserved quantity behind the entanglement (energy, momentum, angular momentum, spin projection, ...) is properly conserved. It appears that Everett did not fully comprehend the central conundrum of nonlocality and entanglement.
>
> The question raised by the many-worlds view is: for two measurements on entangled subsystems made at widely separated sites, how do the outcomes of the two measurements that are consistent with the entanglement conditions end up in the same “world”? One can envision two scenarios: (1) the first measurement to occur instantaneously splits off a world characteristic of a given outcome, and in that world the entanglements conditions restrict the possible outcomes of the second measurement to those consistent with the entanglement; or (2) each measurement creates a split of worlds, the splits somehow propagate (at light-speed?), meet in some intermediate region, and the splits of one measurement join or avoid the splits from the other measurement, linking up so that the entanglement conditions are respected. The problem with (1) is that it is inconsistent with special relativity. In most cases, the choice of which measurement occurs first depends on the reference frame from which the system is viewed. There is no unique “first measurement” that can create a world in which the second measurement can operate. So this approach manifestly violates Lorentz invariance. The problem with (2) is that propagating and self-sorting world splits and the mechanisms behind them are far removed from the formalism of quantum mechanics and from the spirit of Everett’s minimalist approach to the interpretational problem. As far as we can tell, the Everettian program has produced no plausible account of how quantum nonlocality and multiple measurements on entangled systems should be viewed or interpreted.
This paper, which was written in 2010, does not cite let alone address the arguments given in
http://arxiv.org/abs/quant-ph/0104033
http://arxiv.org/abs/quant-ph/9906007
which were written in or before 2001. See also
http://arxiv.org/abs/1109.6223
Whether a particular region of the multiverse is differentiated into parallel universes or not is determined by the flaw of information. Universes are structures within the multiverse in which information flaws freely. Information does not flow from the version of me sitting 1mm to the right of where I am sitting now to the keyboard of the version of my computer on which this is being typed. So the version of me in this universe has typed the words you're reading, the version of me 1mm to my right has not. In each interaction conservation laws are described by the Hamiltonian having certain symmetries. All of this is described in the formalism of quantum mechanics and so can hardly be counted as alien to it.
The authors of the paper choose instead to discuss a popular science article by David in which he argues that the multiverse is needed to explain how quantum computation works. The authors then say (p. 8):
> Deutsch is right: the universe is more than what we see around us, but that does not mean that it has to be a multiverse, in which there are literally actual world counterparts to our own and in which all possible outcomes are actualized. The portion that we do not see, and that is responsible for the power of quantum computing over classical computing, can instead be interpreted as that which is real but not actualized: dynamical possibilities.
What does "real but not actualized" mean? We get a quote from Heisenberg on p. 8 which simply restates the idea without explaining it. The advocates of the TI can't clearly state what this slogan means. If it means that all of the outcomes happen, then their theory is the MWI with unnecessary bells and whistles. If it means that some of of the outcomes don't happen, then they have to explain how stuff that doesn't happen affects stuff that does happen in interference experiments, the EPR experiment and other quantum mechanical phenomena. But as David explained in FoR Chapter 4 the way to decide what's real is to work out what our best explanations say about what exists. QM says that the multiverse exists.
> Yet if current researchers are willing to countenance such admittedly ‘fantastic’ (Kent’s term) or speculative features as bifurcating worlds and observers,
This is very poor. It's a bit rich to say that something is 'fantastic' when you have nothing to offer as an alternative account of how the world works but a load of unexplained philosophical verbiage about "real but not actual".
> ‘probability’ redefined as not requiring uncertainty about outcome,
This would be a viable objection if an alternative was offered that was a good explanation and did feature uncertainty, but no such explanation is offered. Rather the offers bluntly assert the Born rule is true on p.5 because it is the amplitude of the wave for that outcome. Why is the amplitude relevant? No explanation is given. So even if all the decision theoretic arguments were complete failures, the MWI would be no worse off than the TI.
> observer-dependent and ultimately subjective divisions of the world into ‘system’, ‘observer’, and ‘environment’,
Divisions of the world into system and environment are subjective are they? Gosh, we'd better throw out the whole of thermodynamics then. And if the advocates of the TI aren't willing to do that, we can take the objection as exactly as much a refutation of the MWI as it is of thermodynamics: that is, not a refutation at all. It may be something that needs to be elaborated and explained better: that's what science and philosophy are for.
One last objection:
> Note that in his (1998), Deutsch wants to describe such an electron as existing in all his many (interfering) worlds—being actualized in all possible different outcomes in separate worlds. But since no observational basis has been specified, are these many worlds ones in which the electron has a definite (more precisely, narrowly localized) position, i.e. a splitting with respect to the position basis?
This is determined by interactions between systems. Any particular interaction copies information about some observables but not others. Differentiation of the multiverse into parallel universes happens with respect to the observables about which information is transmitted.
Alan
http://arxiv.org/abs/1001.2867
Let's have a look at the substance of their claims.
They describe the EPR experiment. They then say
> The more general consideration of dual or multiple measurements on entangled quantum systems, as in EPR experiments, has not been adequately addressed. The interpretational problem with such systems is that while each individual measurement may have many possible outcomes, because of conservation laws acting in entangled systems, only certain outcomes of a second measurement are permitted for any particular outcome of the first measurement. Only pairs of outcomes are permitted for which the conserved quantity behind the entanglement (energy, momentum, angular momentum, spin projection, ...) is properly conserved. It appears that Everett did not fully comprehend the central conundrum of nonlocality and entanglement.
>
> The question raised by the many-worlds view is: for two measurements on entangled subsystems made at widely separated sites, how do the outcomes of the two measurements that are consistent with the entanglement conditions end up in the same “world”? One can envision two scenarios: (1) the first measurement to occur instantaneously splits off a world characteristic of a given outcome, and in that world the entanglements conditions restrict the possible outcomes of the second measurement to those consistent with the entanglement; or (2) each measurement creates a split of worlds, the splits somehow propagate (at light-speed?), meet in some intermediate region, and the splits of one measurement join or avoid the splits from the other measurement, linking up so that the entanglement conditions are respected. The problem with (1) is that it is inconsistent with special relativity. In most cases, the choice of which measurement occurs first depends on the reference frame from which the system is viewed. There is no unique “first measurement” that can create a world in which the second measurement can operate. So this approach manifestly violates Lorentz invariance. The problem with (2) is that propagating and self-sorting world splits and the mechanisms behind them are far removed from the formalism of quantum mechanics and from the spirit of Everett’s minimalist approach to the interpretational problem. As far as we can tell, the Everettian program has produced no plausible account of how quantum nonlocality and multiple measurements on entangled systems should be viewed or interpreted.
This paper, which was written in 2010, does not cite let alone address the arguments given in
http://arxiv.org/abs/quant-ph/0104033
http://arxiv.org/abs/quant-ph/9906007
which were written in or before 2001. See also
http://arxiv.org/abs/1109.6223
Whether a particular region of the multiverse is differentiated into parallel universes or not is determined by the flaw of information. Universes are structures within the multiverse in which information flaws freely. Information does not flow from the version of me sitting 1mm to the right of where I am sitting now to the keyboard of the version of my computer on which this is being typed. So the version of me in this universe has typed the words you're reading, the version of me 1mm to my right has not. In each interaction conservation laws are described by the Hamiltonian having certain symmetries. All of this is described in the formalism of quantum mechanics and so can hardly be counted as alien to it.
The authors of the paper choose instead to discuss a popular science article by David in which he argues that the multiverse is needed to explain how quantum computation works. The authors then say (p. 8):
> Deutsch is right: the universe is more than what we see around us, but that does not mean that it has to be a multiverse, in which there are literally actual world counterparts to our own and in which all possible outcomes are actualized. The portion that we do not see, and that is responsible for the power of quantum computing over classical computing, can instead be interpreted as that which is real but not actualized: dynamical possibilities.
What does "real but not actualized" mean? We get a quote from Heisenberg on p. 8 which simply restates the idea without explaining it. The advocates of the TI can't clearly state what this slogan means. If it means that all of the outcomes happen, then their theory is the MWI with unnecessary bells and whistles. If it means that some of of the outcomes don't happen, then they have to explain how stuff that doesn't happen affects stuff that does happen in interference experiments, the EPR experiment and other quantum mechanical phenomena. But as David explained in FoR Chapter 4 the way to decide what's real is to work out what our best explanations say about what exists. QM says that the multiverse exists.
> Yet if current researchers are willing to countenance such admittedly ‘fantastic’ (Kent’s term) or speculative features as bifurcating worlds and observers,
This is very poor. It's a bit rich to say that something is 'fantastic' when you have nothing to offer as an alternative account of how the world works but a load of unexplained philosophical verbiage about "real but not actual".
> ‘probability’ redefined as not requiring uncertainty about outcome,
This would be a viable objection if an alternative was offered that was a good explanation and did feature uncertainty, but no such explanation is offered. Rather the offers bluntly assert the Born rule is true on p.5 because it is the amplitude of the wave for that outcome. Why is the amplitude relevant? No explanation is given. So even if all the decision theoretic arguments were complete failures, the MWI would be no worse off than the TI.
> observer-dependent and ultimately subjective divisions of the world into ‘system’, ‘observer’, and ‘environment’,
Divisions of the world into system and environment are subjective are they? Gosh, we'd better throw out the whole of thermodynamics then. And if the advocates of the TI aren't willing to do that, we can take the objection as exactly as much a refutation of the MWI as it is of thermodynamics: that is, not a refutation at all. It may be something that needs to be elaborated and explained better: that's what science and philosophy are for.
One last objection:
> Note that in his (1998), Deutsch wants to describe such an electron as existing in all his many (interfering) worlds—being actualized in all possible different outcomes in separate worlds. But since no observational basis has been specified, are these many worlds ones in which the electron has a definite (more precisely, narrowly localized) position, i.e. a splitting with respect to the position basis?
This is determined by interactions between systems. Any particular interaction copies information about some observables but not others. Differentiation of the multiverse into parallel universes happens with respect to the observables about which information is transmitted.
Alan
Elliot Temple
Jan 20, 2013, 6:28:25 PM1/20/13
to beginning-...@googlegroups.com
On Jan 20, 2013, at 11:39 AM, Alan Forrester <alanmichae...@googlemail.com> wrote:
> What does "real but not actualized" mean?
-- Elliot Temple
http://fallibleideas.com/
reka...@hotmail.com
May 19, 2013, 7:11:30 PM5/19/13
to beginning-...@googlegroups.com, alanmichae...@googlemail.com
Dear Alan and Elliot:
I hope you won't be so quick to dismiss TI and my development of it based on one paper. Actually, Elliot, I really do care very much about the important metaphysical distinction between 'possible' and 'actual,' and I discuss this in detail in my new book from Cambridge U. Press. So I do hope you will give this approach a chance by reading this more detailed material. Info on the book is here: www.cambridge.org/9780521764155, or see my website for preview material and a 20% discount: http://transactionalinterpretation.org
Many thanks for your interest,
R. E. Kastner
I hope you won't be so quick to dismiss TI and my development of it based on one paper. Actually, Elliot, I really do care very much about the important metaphysical distinction between 'possible' and 'actual,' and I discuss this in detail in my new book from Cambridge U. Press. So I do hope you will give this approach a chance by reading this more detailed material. Info on the book is here: www.cambridge.org/9780521764155, or see my website for preview material and a 20% discount: http://transactionalinterpretation.org
Many thanks for your interest,
R. E. Kastner
Reply all
Reply to author
Forward
0 new messages