Copenhagen: Measured or Observed?

[Dmitry67]

For once I agree with most of what you say, but it's wrong to call this "the CI". You're only talking about option 4 on Demystifier's list. (And that's assuming I'm interpreting his 4 correctly).

Sorry, what list?

 

[kote]

I would say that when a photon goes through a polarizing beam splitter (PBS), then I would expect one of the following to occur:

a) the psi function collapses and a whole photon goes through one or the other of the PBS paths;

b) the psi function splits with 50% headed down one path, 50% headed down the other, but collapse does not occur until later when the photon is detected.

I think it is interesting that the 2 x 50% parts - i.e. b) - can later be reconstituted to create a 100% whole again; whereas the a) scenario you would not expect that since the collapse already occurred. So I would consider that as "evidence" for b) - and makes the psi wave relatively real and physical. (The reason I put evidence in quotes is because I know this is not an ironclad argument and is interpretation-dependent.)

I agree that this is exactly what you would expect if you were trying to visualize what's going on during the experiment. Bohr denies that such visualization is possible. You have the phenomenon of a particular experimental reading that depends necessarily on your experimental arrangement and whose description is incomplete without reference to the entire setup. Complementarity is defined by its unvisualizability. Thinking about what the photon does in the meantime, for Bohr/CI, is impossible. You can only talk about the entire setup. We think classically, and until you get that classical system output, everything is simply unintelligible.

 

[Fra]

Hence, the CI description “collapse of the (unreal) wavefunction changes our knowledge about it”

Did you really mean this or is it a typo? ie. knowledge about the wavefunction? (assuming it refers to wavefunction?)

As I see it, the wavefunction supposedly IS the structure for the knowledge. The hilbert space is the context for information to be define, and relative to that context, the wavefunction IS the information (encoded of course).

In this respect I'm with how Zurek put it - what the observer knows is indistinguishable from what the observer is.

The problem is here the assumption when you compose complex systems are products of hilbert spaces with separate wavefunctions. In my view, for a giver observer, there really is only one "system" to interact with and this is nothing less than the entire (unkonwn) environment, period. That's how I see it.

(That's close to Zurek, but I disagree with Zurek on other points.)

/Fredrik

 

[Dmitry67]

Fra, is Observer irreducable and fundametal?
In another words, can we write (in principle) very complicated function IsObserver(X) where X is any system, and IsObserver() returns boolean value?

 

[DrChinese]

I agree that this is exactly what you would expect if you were trying to visualize what's going on during the experiment. Bohr denies that such visualization is possible. ...

And of course now we have all kinds of exotic entanglement to consider, not available in Bohr's day. Strangely (to me), we can draw all kinds of ways to entangle things... but we can't see when or how or even whether collapse occurs.

But we know it when we see it.

 

[RUTA]

And of course now we have all kinds of exotic entanglement to consider, not available in Bohr's day. Strangely (to me), we can draw all kinds of ways to entangle things... but we can't see when or how or even whether collapse occurs.

But we know it when we see it.

What do you think about quantum eraser experiments?

 

[DrChinese]

What do you think about quantum eraser experiments?

Sure, I mean those are perfect examples of how difficult it is to ignore the role of collapse. On the one hand, you can still hold Bohr's view and there is nothing wrong with it. You simply look at the entire setup as one black box. On the other hand, we are talking about some pretty complex and exotic formats. Again, no issue with the formalism, just saying that it gets harder and harder not to imagine that you have to draw a line somewhere - and we can't seem to figure out where.

 

[Fredrik]

Yes, but CI proponents would say "well, the first measurement was not a real one!" :)

Or they would distinguish between "state preparation" and "measurement". A Stern-Gerlach magnet just prepares the system in a certain state.

Sorry, what list?

This one.

I agree with him. There's no clear set of axioms that defines "the CI". There is also no set of axioms that defines "the MWI", or any other "interpretation". (I think Bohm is an exception, but I don't know enough about it to say for sure). That's why I think almost everything that's been written about interpretations is meaningless nonsense.

 

[kote]

I agree with him. There's no clear set of axioms that defines "the CI". There is also no set of axioms that defines "the MWI", or any other "interpretation". (I think Bohm is an exception, but I don't know enough about it to say for sure). That's why I think almost everything that's been written about interpretations is meaningless nonsense.

Bohm is certainly not an exception . Well... let me clarify. David Bohm's view of quantum mechanics was very different from what we now call Bohmian Mechanics. Bohm himself had a relatively consistent (although organically developed) interpretation. Within the Bohmian Interpretation (as opposed to Bohm's Interpretation) there are still variations.

I think interpretations are meaningful, but they have been very confused. I think it's correct to say that there is no single CI in its present usage. There is certainly one consistent interpretation from Bohr though, which can be very different from what is called CI, similarly to how the Bohmian Interpretation is not Bohm's Interpretation.

 

[Fra]

Fra, is Observer irreducable and fundametal?
In another words, can we write (in principle) very complicated function IsObserver(X) where X is any system, and IsObserver() returns boolean value?

Ok I see. You wonder if there is something special about systems to qualify as observers?

IMO, no. Any system is an observer.

However not every _conceivable imagined_ system is a FIT - but then that's consistent with observations because matter is very structured (obeying symmetries etc, so it's not anything goes). Only fit observers are observed, like in nature there is a reason for why only particular speices populate the world.

So on one hand I'd say IsObserer(X) = true for any X. But I think the interesting question is which particular systems are selected to populate the universe.

It's like considering a mathematical imaginary theory were all kinds of twisted particles may possibly exist, and then ask why don't they? Here fitness and evolution of law comes in. IMO physical law is encoded in the microstructure of the systems that does populate the universe.

So evolution of law is the other side of the coin to evolution of physical systems.

The answer lies in understanding the evolutionary process, like Darwins theory in biology. Each environment selects it's preferred inhabitants, the circularity lies in the fact that the environment IS a collection of many inhabitants. But this understanding is not worked out yet. Solve that problem and we should see some answers to what constrains not just observers but ANY physical system, which is basically the question for the structure of matter I believe.

/Fredrik

 

[Fra]

Fra, is Observer irreducable and fundametal?
In another words, can we write (in principle) very complicated function IsObserver(X) where X is any system, and IsObserver() returns boolean value?

The way I envision this program - which I think is the only consistent way I know of - is that the reconstruction of a probability theory(*) I imagine will also be a reconstruction of the makeup of viable physical systems, and thus observers.

The physical systems in nature are not arbitrary, as we know from the standard model. So indeed "any system" is already constrained if we count only those that we see in nature, and the constraints of "any system" is the same that qualifies them to be a "physical observer". So given that "the system" is constructed in a physical process (and not just beeing made up in some wild theory) then all physical systems qualifis as observers. That's my view.

(*) Since I share part of the spirit of Jaynes and Ariel, I think that somehow probability theory is an extension to logic and indeed the logic of science and reasning, but only sort of! The current continuum probability makes no sense. So the reconstruction I imagein, is really a reconstruction of a more physical measure theory, and in my interpretation and observer or a system IS thought of as a measure-complex. This measure complex is a generalisation and replacement of the hilbert space abstraction. Actually I expect there to be a one to one mapping from one measure-complex to one hilbert space, but the problem is that hilbert spaaces are normally thought of as qualifiying context. In my view the measure-complexes are embodied by observers, and these are evolving. So in my view hilbert spaces are evolving - which is why I find hilbert space the first axioms of ordinary QM to be inappropriate.

So I am working on an extension/reconstruction of probability theory which is the basis for statistical physics and classical thermodynamics. The new "information-complex" theory will (this is my conjecture) like classical probability leads to thermodynamics, lead to ordinary physics with much more complex interactions. The classical state space corresponding to a probability space, is replaced by a measure-complex (a system of related discrete spaces) which in a differential sense corresponds to hilbert spaces.

This plan, implies also that the actions or hamiltonians are encoded in the measure-complexes simply as natural diffusion type actions. The measure complex vs simple probability space is what I expect will make the interaction nontrivial and diverse, rather than just diffusion type like in thermodynamics.

/Fredrik

 

[Fra]

B – consciousness can be explained in physical terms, hence, it is just a state of a physical system. See above.

Given your choices, B is my closest match. (if I replace conscioussness with observer here)

But if we say, an observer is just a physical system, then the question how an observer works, amounts of question how a physical system works - ultimately questioning matter and it's physical action pattern.

/Fredrik

 

[Fra]

But this is no clean CI. I sure don't rely on a classical reality in my view. I rely on the only think I can rely on - my prior structure and state. So, my actions are as if, this was classical reality. But my actions are (in the differential sense) indifferent to wether my view matches those of my fellow neighbouring systems. If not, I will be informed about this from feedback from my environment. One step at a time. To take one step, I do not need to know what's 100 steps ahead. Only my expectation guides me, and I think the same applies to physical actions. This is a form of locality of information. Only information at hand, influences your action. This does however not say much about the future. By I think evolution works from the present.

/Fredrik

 

[Fra]

How's this derivation of the CI?

Zurek, http://arxiv.org/abs/0707.2832 : "Finally, we point out that monitoring of the system by the environment (process responsible for decoherence) will typically leave behind multiple copies of its pointer states. Only states that can survive decoherence can produce information theoretic progeny in this manner."

There's too much to object to there in there. The axioms he start with about the hilbert spaces are unacceptable starting points. This represents a nontrivial baggage. So instead I'll note what I like.

- On his idea of the environment as an information pool, in which copies of an interacting system is propagated and reproduced is I think very good. This conceptually makes excellent sense. I'd argue thta this applies not only to copies of states, but also copies of the action, that the environment also evolves a consistency with the systems action/interaction properties, and thta this can be thought of as the DNA of physical law. This could be extended to an alternative way of implementing evolving law, that is different from smolins CNS. This is then very close to my thinking.

So what I mean is that by interacting with the environment, you're not merely spreading your word (ie copying your state) you are also spreading your REASONING, ie copying your action, and this is close to the DNA of physical law (which would be symmetries in the action), and this can explain how evolution of law works, in a quite more interesting way than the CNS.

An environment that is equilibrated as per a particular symmetry, would then by selection favour growth of "consistent" systems.

My main objection is that a lot of the premises, used to make the inference Zurek does, is IMO no different than other information. So I think Zurek is still fairly conservative here. I think we need to be more radical to gain more predictive power.

/Fredrik