QM says no observer, no existence

[vanesch]

So would it be impossible to somehow "mind-meld" consistent histories with MWI, or with Everett's relative state? Postulate that every consciousness' past light cone be self consistent?

Yes, that's what I mean: this can be solved again by postulate, but this time it looks like a plaster! It would mean somehow that at its first association with a body state, the consciousness would be assigned to that bodystate according to the Born rule too. My original idea was in fact that EACH body state that was "receptive" of a consciousness, would get one :-) In that consciousness would then be an "emerging property" of a state with the right structure. But that cannot be the case, because MOST bodystates are in a branch with total absence of a Born rule. You could argue that sometimes the Born rule wasn't followed (after all, we ARE the product of a highly singular history!), but that's not good enough: most bodystates live in branches where there is not the slightest hint of a Born rule in their past. So no "experimenter" in that branch would have gotten the idea that there was such a thing as the Born rule, and physics books would be different in that branch.
Which leads to the conclusion that most body states that would be around in different branches cannot be conscious ; only those that are in "long" branches (with high Hilbert norm).

 

[El Hombre Invisible]

Patrick

Thanks again for your replies and willingness to continue this discussion with such a novice. Rather than reply to individual parts of your posts and risk overlapping points, I'll try to extract the main areas of interest and treat them separately.

1. Schrodinger's Lucky Dip

The apparatus is set up in such a way that the photon could end up in anyone of n photon detectors. Each photon detector is connect to a machine that pulls a particular item out of one box and places it in a hitherto empty one. The entire apparatus is enclosed and only the initially empty box is accessible. We run the experiment and open the box. Now, by your logic, the quantum system inside that box is in a superposition of |cat> + |dog> + |blanket> + |banana> + |monkey_with_a_hat> + |brick> + |bonsai tree> + ...

Maybe you yourself are in a superposition of |conscious_human> + |conscious_banana>, and only your consciousness observes you as human. All this time people have been telling you that you are a banana, but you only observe them telling you you're human.

Then you know you need a holiday.


2. Macroscopic properties of quantum systems

If I follow you correctly, you hold the view that if a macroscopic system is a quantum system (with which I agree) then it must have macroscopic properties that are also in a superposition of states. I think the Schrodinger's Cat (or Lucky Dip) experiment clouds this issue because it is concerned with macroscopic properties determined by microscopic wave functions.

Instead, take the double-slit electron experiment again. The macroscopic system I'm interested in is the screen, and the macroscopic property I'm interested in is the interference pattern. If this must be in superposition, then it's wave function will include |interference_pattern> + |no_interference_pattern>. I run the experiment with no light source at the slits, so nothing that could cause decoherence. The screen's property is in superposition, and it is only our consciousness that observes the one term |interference_pattern>. So now we have a paradox. The property is in superposition, and that includes the term |no_interference>, in spite of the very phenomenon that leads us to believe in the superposition of macroscopic states in the first place.

If you accept that, when no scattering takes place, the macroscopic property (not state) of the screen that I'm interested in is NOT in superposition, then it follows that macroscopic properties in general DO NOT have to be in superposition. The system (screen) itself is still in a superposition of states, but all of these give you the same macroscopic property.

Then, going back to Schrodinger's Cat, the question is not how a quantum system can seem to have macroscopic properties (i.e. observable properties) not in superposition, but how the wave function evolved such that the system's superposition results in that macroscopic property. This is not in violation of the any of the unitary QM assumptions you mentioned, and doesn't rely on curious reinterpretations of consciousness and conscious awareness.

Emergence, I think, is the key to differentiating between properties that may and may not be in a superposition of properties. The more elementary the system is, the more it will appear in superposition. But for larger, more complex systems, the individual wave functions of each elementary particle within it get obfuscated by the system as a whole - a 'can't see the forest for the trees' situation. States such as |live_cat> and |dead_cat> are not contained within the protons, neutrons and electrons (which do have superposition of properties) that comprise the cat, not even the atoms, or even individual molecules, not even cells, muscles, organs and limbs. It takes a whole cat for the |live_cat> and |dead_cat> states to emerge. The cat is in a superposition of states, but none of these include |dead_cat> or |live_cat>, and the wave function of a given single electron, say, won't make a difference to that property.


3. Decoherence in macroscopic systems

You've corrected me on my employment of the term decoherence, but this I think has more to do with my loose grasp of the terminology of QM and very little understanding of how to get from A to B when talking about decoherence affecting macroscopic systems. I know what I mean, but cannot express it so I'm going to rely on your superior intelligence and knowledge to interpret my gibberish.

Wikipedia says of decoherence:

"The effect is basically one in which the system under consideration loses the phase coherence between certain components of its quantum mechanical state and hence no longer exhibits the essentially quantum properties (such as superposition and entanglement) associated with such coherence."

This still doesn't tell me why, but this is what I was referring to - losing coherence related to loss of or change to superposition.


4. Definition of consciousness

Your definition of consciousness renders the question of whether another quantum system is conscious or not irrelevant. However, it does not support the "the assumption that the superposition principle is universally valid" or "the fact that we do not observe the world in a superposition", so far as I can tell. The assumptions of unitary QM cannot proceed if I claim to be consciously aware of all systems in superpositioned states. Likewise you cannot proceed with the assumption that, say, a photon detector is 'aware' of both 'detected' and 'not detected' since we do not know whether or not it is conscious. (Here I assume there to be such a thing as non-conscious awareness, e.g. the man in the Chinese room.) It is the 'we' part of "the fact that we do not observe the world in a superposition" that demands a definition of consciousness. Who is 'we', and what criteria must they fulfill to be a part of that collective?


5. Interactions between consciousness and the physical world

If consciousness is some kind of passive observer that cannot alter the wave function of a physical system, then its existence or non-existence should be irrelevant. How, then, can this post, and the many articles and fields of science preoccupied with it, actually exist?

An interesting possible counter-example to your argument is suicide. No animal that is blissfully unaware of its own existence can know to end it. It cannot be described behaviourally since you can only do it once (Pavlovian suicide just would not work). Humans alone are known to do this, and humans alone are known for sure to have consciousness. If consciousness is required to commit suicide, then a more emphatic example of consciousness altering a wave function you'd be hard pressed to find.

Lastly on this point, from what I've read of MWI (which is admittedly little), it does not necessitate a non-physical consciousness associated to one term of a wave function. If consciousness does get entangled with the system, say the cat experiment, then you have:

|photon_reflected>|photon_undetected>|live_cat>|consciously_observe_live_cat> + |photon_not_reflected>|photon_detected>|dead_cat>|consciously_observe_dead_cat>

Here, two worlds exist in superposition, and as my consciousness is entangled with the system, we will observe the term of whichever branch we are in. What's the problem with this?


6. Nature of consciousness and conscious observation

I think your claim that we cannot seek to explain consciousness as a physical emergent property is at best premature and at worst denial, and building in a fail-safe that says any physical explanation for consciousness is by default defining consciousness as something else only highlights your own lack of definition. And while you are quite correct to introduce the distinction between consciousness and behaviour, I cannot agree that they are entirely exclusive.

Your example of a computer that expresses pain when you press the space bar is really a rewording of the Chinese room argument, which would also illustrate your point about behaviour very well. You put a man who doesn't speak Chinese in a room with a set of instructions on what to do when given various Chinese symbols. You feed symbols into the room, he applies the rules, and you get some converted symbols out of the room. This may well be entirely indistingiushable from a room with an actual Chinese person in it who knows the meaning of the symbols he receives and of those he sends out. From outside the rooms, the two may appear indistinguishable. The former is analogous with a computer (and in fact was proposed as an illustration of why a computer can not be considered a 'thinking machine' no matter how sophisticated, complex and rigorous the instructions). But the key difference is that the second man knows WHY he outputs what he does. And of course, the first man is not going to output the symbols for 'MY GOD, THIS IS BORING' unless you specifically instruct him to.

This notion of 'why' is the important detail when distinguishing between conscious and non-conscious systems, or even conscious and non-conscious reactions within a conscious system. Subliminal imagery, for instance, causes unconscious reactions. A single frame of a cheeseburger spliced into a film may make the viewer go and hit McDonalds afterwards, but without any knowledge of WHY they did so. A more overt image may yield a different reaction: the viewer WANTS to buy a cheeseburger, but also know WHY they want to and as a result overrides the unconscious reaction and does not buy any food at all. If the longer image had been shown to a physically identical viewer incapable of that conscious objection, McDonalds may still be 69 pence up on the deal. This, and indeed any other type of conscious decision, would contradict your theory that consciousness cannot influence a physical system - only observe it.

Tests such as the spot experiment are equivilent to trying to devise some test of the Chinese rooms that a conscious occupant would pass and a non-conscious one fail or vice versa, i.e. distinguish between them by observing behaviour unique to conscious beings. It will fail if the behaviour being studied is NOT unique to conscious beings, in which case the hunt for a new behaviour will start. Another way would be to just enter the room and determine for yourself, and this is equivilent to more physiological studies of consciousness. Neither give us anything concrete, but nonetheless they have told us a lot along the way. This is why I find your assertion that consciousness is a non-physical phenomenon but, were it a physical phenomenon, we would not find a physical explanation for it a stretch. Certainly making such fantastic claims and curious assumptions regarding a field in such an infant stage but nonetheless progressing well is really putting all your eggs in one basket.


Overall, your definition of consciousness made general and description of how it interacts with the physical world (passive observer) may well partially address the question: "why, if all physical systems are in a state of superposition, do we not observe this?", which we may otherwise have some hope of answering physically. However, all it really does is push the answer out of the physical world where we may hope to fully grasp it and into some non-physical realm where we have no hope of grasping it. On top of that, it introduces countless questions about the nature of this non-physical existence and how it interfaces with the physical universe that also cannot be answered. Any question about this aspect of nature is going to end in 'we cannot seek to understand, only to accept' which sounds all too familiar.

 

[vanesch]

Maybe you yourself are in a superposition of |conscious_human> + |conscious_banana>, and only your consciousness observes you as human. All this time people have been telling you that you are a banana, but you only observe them telling you you're human.

Then you know you need a holiday.

I may need a holiday, but I still maintain that there is no way of avoiding the above conclusion when you strictly apply unitary QM, which is at the origin of all interpretational issues.

2. Macroscopic properties of quantum systems

If I follow you correctly, you hold the view that if a macroscopic system is a quantum system (with which I agree) then it must have macroscopic properties that are also in a superposition of states. I think the Schrodinger's Cat (or Lucky Dip) experiment clouds this issue because it is concerned with macroscopic properties determined by microscopic wave functions.

I don't see how it clouds this: it illustrates the issue perfectly, I think - and Schroedinger was the first one to realize this. His argument still stands.

Instead, take the double-slit electron experiment again. The macroscopic system I'm interested in is the screen, and the macroscopic property I'm interested in is the interference pattern. If this must be in superposition, then it's wave function will include |interference_pattern> + |no_interference_pattern>. I run the experiment with no light source at the slits, so nothing that could cause decoherence. The screen's property is in superposition, and it is only our consciousness that observes the one term |interference_pattern>. So now we have a paradox. The property is in superposition, and that includes the term |no_interference>, in spite of the very phenomenon that leads us to believe in the superposition of macroscopic states in the first place.

Ok, let us debunk this argument. If the macroscopic property you're interested in is "interference pattern or not" (which is the outcome of a perfectly respectable observable with two eigenvalues, namely 1 and 0, 1 if you have an interference pattern, and 0 if you don't), you've simply split the Hilbert space into a sum of two subspaces, each corresponding to the eigenspaces of the respective eigenvalues, 1 and 0.

However your measurement system is NOT the screen you're thinking about: the screen does NOT measure "interference or no interference" ! It measures "impact position of the electron". Indeed, for a single electron, you do not have an outcome which tells you "interference" or "no interference". You get a single position.
But you CAN find out a device that observes interference or not: for instance a diffraction grating, in which an interference pattern will be diffracted into a certain direction (where you can put a detector), and a non-interference pattern will end up somewhere else. THIS is now the measurement interaction that is needed to amplify the state "interference" or not to the macroscopic level.
Now, if your original electron beam is hitting the two holes at once, you simply are in a state that will evolve in a pure "interference" state. You are measuring an eigenstate of the system, and you will get your outcome with certainty. This is the case when you go from microscopic |a> to macroscopic |A>. You have a pure pointer state (which, here, is "interference pattern"). So in this case, your screen is NOT IN A SUPERPOSITION IN THE BASIS IN WHICH YOU MEASURE (and which is amplified to the macroscopic level). No "branching" occurs here.
When you switch on your light source, however, the microscopic state of the electrons changes (due to the interaction with the light source), and your incident state is now indeed a superposition of "interference" and "no interference". Your diffraction grating will now give you a superposition of a click in the "detector" and not.

If you accept that, when no scattering takes place, the macroscopic property (not state) of the screen that I'm interested in is NOT in superposition, then it follows that macroscopic properties in general DO NOT have to be in superposition.

Unfortunately I don't accept that, because "interference" (of a single electron) is not what a screen measures (is not the basis in which it decoheres, if you want to). A screen couples to "position of inpact".

Then, going back to Schrodinger's Cat, the question is not how a quantum system can seem to have macroscopic properties (i.e. observable properties) not in superposition, but how the wave function evolved such that the system's superposition results in that macroscopic property. This is not in violation of the any of the unitary QM assumptions you mentioned, and doesn't rely on curious reinterpretations of consciousness and conscious awareness.

Well, you can try to think up a macroscopic property that corresponds to the wavefunction of a cat being in a superposition of a lot of its atoms of it "here in the box" and a lot of its atoms being in an orbit around the earth, and you'll have a hard time finding such a property! The fact that this superposition is part of its wavefunction, again, follows simply from the linearity of the time evolution operator. I have the impression that this argument didn't sink in:
if microstate |a> leads to macrostate |A> for sure,
(say, if the photon hits the left of your screen, a bomb goes off in London)
and if microstate |b> leads to macrostate |B> for sure (say, a bomb goes off in Paris)
then, by linearity, a microstate in a superposition:
u|a> + b|b> leads to the macrostate u|A> + v|B> for sure.

Now you can argue all day, if the time evolution operator is linear, you cannot escape that. And "a bomb goes of in London" is macroscopically different from "a bomb goes off in Paris" for sure, so you will have to explain me what happens with this strange state u|A> + v|B>. To what macroscopic property it leads.

Emergence, I think, is the key to differentiating between properties that may and may not be in a superposition of properties.
[...]
The cat is in a superposition of states, but none of these include |dead_cat> or |live_cat>, and the wave function of a given single electron, say, won't make a difference to that property.

I agree with that, except the last sentence. Of course, "dead cat" is not a single state, it is a whole subspace of states of the particles of the cat, which correspond to "the cat is dead". It can be poisened, crushed, squeezed, whatever. There are many states that correspond to a dead cat. "dead cat" simply refers to the entire subspace of these states. So WITHIN this subspace, you are right, the dead cat can be in any of these states, and their superpositions, and it is still a dead cat.
Idem for the live cat (probably a smaller space: there are less ways for the atoms of a cat to make up a live cat than a dead one :-).
However, that's not the point. The point is that if the time evolution is such, that a microstate |a> will evolve the cat state into the subspace of "live cat" somewhere, but for sure, and the same time evolution operator will evolve from microstate |b> into a cat state in the subspace of "dead cat" somewhere, the linearity of that time evolution operator will necessarily evolve a microstate u|a> + v|b> the cat in a superposition of a state in the "dead" cat state and the "live cat" state WITH EXACTLY THE SAME COEFFICIENTS u and v. That's the very definition of linearity.
What you say however, is correct, that these states are very complex and that it will be extremely difficult to "rotate the observation basis" to observe interference effects showing that we have a genuine superposition. That's what decoherence tells us: that these interference terms are hopelessly suppressed. But that the two terms with coefficients u and v are still present, is unavoidable if you accept unitary (and hence linear) time evolution.

Wikipedia says of decoherence:

"The effect is basically one in which the system under consideration loses the phase coherence between certain components of its quantum mechanical state and hence no longer exhibits the essentially quantum properties (such as superposition and entanglement) associated with such coherence."

I'm affraid that that's an error in Wikipedia, then. The only thing that phase coherence loss induces is the loss of observable interference effects. In fact, decoherence is based upon hopelessly STRONG entanglement!

This still doesn't tell me why, but this is what I was referring to - losing coherence related to loss of or change to superposition.

It is indeed unfortunate that they made exactly this error. Half of the people in the decoherence world think that too, btw, but they are sometimes corrected by people such as Joos and Zeh :-)

It is the 'we' part of "the fact that we do not observe the world in a superposition" that demands a definition of consciousness. Who is 'we', and what criteria must they fulfill to be a part of that collective?

Your observation is correct. It is simply the "we" part. If we'd observe the world in superposition, we'd not have to delve in all this consciousness stuff.
BTW, to be member of the 'we' club...It was majestatis pluralis. *I* is sufficient

 

[vanesch]

(following up ; the post was too long to be submitted, so this is its second part).

If consciousness is some kind of passive observer that cannot alter the wave function of a physical system, then its existence or non-existence should be irrelevant. How, then, can this post, and the many articles and fields of science preoccupied with it, actually exist?

Because non-self aware systems can behave as if they are, and try to solve the problems associated with it. Maybe it is a prerequisite for a non-self aware physical system to be potentially conscious, that it behaves as if it were conscious... (hmmm, I *really* need a holliday).

An interesting possible counter-example to your argument is suicide. No animal that is blissfully unaware of its own existence can know to end it. It cannot be described behaviourally since you can only do it once (Pavlovian suicide just would not work). Humans alone are known to do this, and humans alone are known for sure to have consciousness. If consciousness is required to commit suicide, then a more emphatic example of consciousness altering a wave function you'd be hard pressed to find.

There is a marvelously funny article on that theme on the arxiv. I don't have it here but do a search on "quantum suicide" !

BTW, humans are not known to be conscious. There's only one human I know it is, that's myself. I inductively ASSUME that others are.
But there can be evolutionary reasons to educate species to (sparely) kill themselves, if that inhances the survival chances of near kin (hence, increasing the genetic transfer to the next generation).
Lemmings kill themselves too, no ?

Lastly on this point, from what I've read of MWI (which is admittedly little), it does not necessitate a non-physical consciousness associated to one term of a wave function. If consciousness does get entangled with the system, say the cat experiment, then you have:

|photon_reflected>|photon_undetected>|live_cat>|consciously_observe_live_cat> + |photon_not_reflected>|photon_detected>|dead_cat>|consciously_observe_dead_cat>

Here, two worlds exist in superposition, and as my consciousness is entangled with the system, we will observe the term of whichever branch we are in. What's the problem with this?

Well, if it is one and the same consciousness, you should be aware of BOTH situations, which obviously is not the case, is it ?
But the point of consciousness not entangling in the wavefunction can be easily explained: it is not a physical degree of freedom ! It is not the x-position of a particle or so, or the value of an electric field. Only physical degrees of freedom can get entangled (have their own hilbert space).
If it is something that is ASSOCIATED to a state, then it is not a state itself.

6. Nature of consciousness and conscious observation

I think your claim that we cannot seek to explain consciousness as a physical emergent property is at best premature and at worst denial, and building in a fail-safe that says any physical explanation for consciousness is by default defining consciousness as something else only highlights your own lack of definition.

Well, I thought that that was a philosophically accepted position (in the sense that solipsism is unfalsifiable).

Let me tell you what I really think: I don't really believe that "consciousness bull" myself. I'm only trying to make sense of quantum theory, without altering one iota to the existing formalism (because no such change is in sight).
Now, from the very fact that human bodies end up entangled in wildly different macroscopic states (really, you don't believe me, but it is an elementary property, first noticed by Schroedinger, and of which I don't know how to escape - if someone knows, please tell me), clearly I don't consciously observe that. From there follows almost immediately the dissociation of consciousness from the body as a degree of freedom (maybe not as a STATE). Photodetectors can happily be in superposition, because I'm not aware of them NOT being aware of their world in a superposition. It is just that *I* don't see that.

My real hope resides in that the unitary evolution is just an approximation to something that leads objectively to a collapsed state ; only, this screws up quantum theory completely. There's nothing seriously in view, and people working in superstring theory seem to accept strictly unitary quantum theory far beyond the realm of human bodies for instance. So if my hope is correct, string theory goes into the dustbin. As such, I temper my hopes and try to make sense of strictly unitary QM. You cannot avoid an MWI scenario in that case - except in one view of things, which is strictly epistemological. Then, there is no problem at all: the wavefunction somehow just describes what I know ; if that knowledge changes, it changes. The problem with that is that we now have no ontological description of the world anymore, only what we know ; not even about the world, because the world has no objective description. Pretty poor for a physical theory, I'd say.
Another difficulty is that it is strange that what we are supposed to know about the hydrogen atom or not determines its chemical and physical properties. In classical statistical mechanics, the probability density in phase space is an epistemological state description: the "system" is not in that state, but our knowledge of the system is. Only, in classical statistical mechanics, this epistemological description doesn't influence the dynamics of the system under study (which can be assumed to be in one of the phase space points), it just FOLLOWS the dynamics of the individual phase space points. But in QM, this is not the case: whether or not I 'know' that the particle is confined to a certain region, CHANGES its dynamics ! Knowing stuff about the particle can make it act DIFFERENTLY than not knowing stuff about it.

Your example of a computer that expresses pain when you press the space bar is really a rewording of the Chinese room argument,
[...]
But the key difference is that the second man knows WHY he outputs what he does. And of course, the first man is not going to output the symbols for 'MY GOD, THIS IS BORING' unless you specifically instruct him to.

That's what I meant about artificial intelligence people: they confound "intelligence" (in this case: the ability to dialogue in Chinese) with consciousness. The room has a certain form of intelligence (ability to solve problems): it can "talk" in Chinese.
If the man is inside, however, and remains there for years, he might start to really LEARN Chinese, with all the examples he has been treating, and utter finally his message with a few grammatical errors "my god this is boring, let me out!"

A more overt image may yield a different reaction: the viewer WANTS to buy a cheeseburger, but also know WHY they want to and as a result overrides the unconscious reaction and does not buy any food at all. If the longer image had been shown to a physically identical viewer incapable of that conscious objection, McDonalds may still be 69 pence up on the deal.

That's not necessarily the case. If you'd study in detail the brain and nerve reactions, you should in principle be able to completely predict the fact that if the picture of the hamburger is viewed long enough, this triggers neurons where associations are made with previous experience of eating such a hamburger, and which finally trigger the decision to go to McDonalds, eating one. This is then exactly the same as the computer scientist telling you: don't worry, your computer doesn't feel pain, look, here, is the line in the code where it says: "on space-bar event do" and look, here is the instruction to print on the screen "aauww, you hurt me".
So if you have a physical-chemical explanation of the behaviour of the human body (namely when seeing a picture of a hamburger, you go out to McDonalds eating one in this particular case), there is no "need" for consciousness to take that decision: the physics of the brain determine that. You simply passively EXPERIENCE that, and you THINK you made that decision, but you just experienced your brain physically acting in such a way that the decision was made. A physical reaction (which could be analysed by neurologists, in the same way as computer scientists could read the code of a computer) following the stimuli (and maybe including random generators) made your body buy a hamburger, and your consciousness was just given the feeling it decided it all by itself :-) You know, like with elections.

This, and indeed any other type of conscious decision, would contradict your theory that consciousness cannot influence a physical system - only observe it.

Not if you accept that the decisions are physically decided (which you could find out if only you knew enough of the workings of the brain), and you just "observe yourself deciding".

Tests such as the spot experiment are equivilent to trying to devise some test of the Chinese rooms that a conscious occupant would pass and a non-conscious one fail or vice versa, i.e. distinguish between them by observing behaviour unique to conscious beings. It will fail if the behaviour being studied is NOT unique to conscious beings

This is circular of course. In order to find out if something is conscious, we apply tests which qualify if they only are successfully performed by conscious beings... how do you do the "calibration" ??
For instance, with my claim that granite blocks of more than 2 kg are conscious, and they feel pain when their crystals are broken, how do you find that out with such a test ?
How do we find out if computers are conscious or not ?

Certainly making such fantastic claims and curious assumptions regarding a field in such an infant stage but nonetheless progressing well is really putting all your eggs in one basket.

That's not true, thinking about consciousness is as old as the world :-)
You have only two options:

1) or every observable phenomenon is reducible to the laws of physics (and chemistry and ...) ; in which case, clearly the behaviour of humans is just as well reducible to the laws of physics - it is something I adhere to. Now (and that's what many people, like neurologists and artificial intelligence people do), you can single out certain aspects of human behaviour, and CALL them "conscious behaviour", and then go looking for the physics behind that. But that is - as I said before - missing the point: you've just defined a set of rules of behaviour to be "conscious", in a quite arbitrary way. Of course you'll find the physical reasons for that behaviour. And I'm pretty sure that you will be able to make machines displaying exactly the same behaviour (like CCD cameras, or computers). Ooops, you say, yes, but that's cheating, let's add "THIS" behaviour... ok, a few months later I simulate again that behaviour with a computer... So singling out specific behaviour of physical structures, of which we can know, according to the laws of physics, WHY the structure behaves that way, is not going to help us.

2) Conscious beings fundamentally behave in such a way that cannot be explained by physics, chemistry etc... I don't adhere to it, but if you do, you have then to admit that consciousness is something supernatural which has measurable properties. I don't think it is a viewpoint accepted by most scientists.

However, all it really does is push the answer out of the physical world where we may hope to fully grasp it and into some non-physical realm where we have no hope of grasping it. On top of that, it introduces countless questions about the nature of this non-physical existence and how it interfaces with the physical universe that also cannot be answered. Any question about this aspect of nature is going to end in 'we cannot seek to understand, only to accept' which sounds all too familiar.

I agree with that, it is a "last resort" solution to the almost insolvable riddle in QM of the superposition of body states. But I think you give "too much structure" to consciousness ! If you just call consciousness "what I observe" that will do, no ? It doesn't have to be a blue angle with yellow little wings !
In formulas, it is just an asterix that goes with one term in the state vector decomposition. Can you not accept that on the same level as: "and the state of the universe is a vector in Hilbert space" ?

 

[El Hombre Invisible]

I may need a holiday, but I still maintain that there is no way of avoiding the above conclusion when you strictly apply unitary QM, which is at the origin of all interpretational issues.

You know that was just a bit of light relief, I hope. Although there was a point there too.

I don't see how it clouds this: it illustrates the issue perfectly, I think - and Schroedinger was the first one to realize this. His argument still stands.

What I meant was, the question of whether a macroscopic quantum system has macroscopic properties that have to be in superposition to conform to one of the assumptions you put forth, if true, will not depend on non-predictable outcomes, so we should start simple. But the point I was trying to make was that macroscopic properties should not have to be in superposition.

Ok, let us debunk this argument. If the macroscopic property you're interested in is "interference pattern or not" (which is the outcome of a perfectly respectable observable with two eigenvalues, namely 1 and 0, 1 if you have an interference pattern, and 0 if you don't), you've simply split the Hilbert space into a sum of two subspaces, each corresponding to the eigenspaces of the respective eigenvalues, 1 and 0.

This is what I meant. You're saying, if I understand you correctly, that there exists a property of 'no interference' in superposition with the 'interference' one. But these are both a result of the same experiment. So in the branch containing the the 'no interference' pattern you have a world in which the electron two-slit experiment does not cause interference patterns, which means that the de Broglie waves of the electrons did not interfere, which means that electrons are not in superposition, which means that the assumption that quantum systems are in superposition is false, which means that the property is not in superposition. The best you could hope for is that, at some time in the past, the wave function of the universe evolved into 'electrons interfere' and 'electrons don't interfere' and that, due to consistent histories, we've been in the 'electrons interfere' branch ever since. Since there's no reason to believe the screen even exists in the other branch, the property of the screen I'm interested in will be the same in every branch where it exists, and so not in superposition (another property, though: 'screen_exists'/'screen doesn't exist' may be in superposition).

However your measurement system is NOT the screen you're thinking about: the screen does NOT measure "interference or no interference" ! It measures "impact position of the electron". Indeed, for a single electron, you do not have an outcome which tells you "interference" or "no interference". You get a single position.

That's not what I said, and certainly not what I meant. I am the measuring system, and the thing I'm measuring is the screen - NOT the electrons. The screen has a macroscopic property of an interference pattern - the microscopic properties of where each individual electron landed are not of interest to me. This is as valid a macroscopic property as 'dead cat'/'live cat'.

Now, if your original electron beam is hitting the two holes at once, you simply are in a state that will evolve in a pure "interference" state. You are measuring an eigenstate of the system, and you will get your outcome with certainty. This is the case when you go from microscopic |a> to macroscopic |A>. You have a pure pointer state (which, here, is "interference pattern"). So in this case, your screen is NOT IN A SUPERPOSITION IN THE BASIS IN WHICH YOU MEASURE (and which is amplified to the macroscopic level). No "branching" occurs here.

Good - this was my point. That is, the screen itself, as a quantum system, is still in a superposition of all the possible electron positions and momenta (which will effect the system as long the as momentum is transferred), but it's macroscopic property of 'interference pattern'/'no interference pattern' is NOT in superposition. So we can have quantum systems without macroscopic properties in superposition.

Unfortunately I don't accept that, because "interference" (of a single electron) is not what a screen measures (is not the basis in which it decoheres, if you want to). A screen couples to "position of inpact".

Like I said, we got our wires crossed because this is not what I meant. I've clearly given the impression I'm even dumber than I am. \o/

I have the impression that this argument didn't sink in:
if microstate |a> leads to macrostate |A> for sure,
(say, if the photon hits the left of your screen, a bomb goes off in London)
and if microstate |b> leads to macrostate |B> for sure (say, a bomb goes off in Paris)
then, by linearity, a microstate in a superposition:
u|a> + b|b> leads to the macrostate u|A> + v|B> for sure.

Now you can argue all day, if the time evolution operator is linear, you cannot escape that. And "a bomb goes of in London" is macroscopically different from "a bomb goes off in Paris" for sure, so you will have to explain me what happens with this strange state u|A> + v|B>. To what macroscopic property it leads.

I understand the linearity argument. That's kind of my what my point was about earlier. I don't think microscopic states lead to macroscopic states for sure, because a macroscopic system is so much more complex than, say, a fundamental particle.

In one article of decoherence I read (which holds a completely different view to yours) it stated that:
1. decoherence evolves incredibly rapidly, such that the entire Universe's state is resolved quickly;
2. decoherence occurs constantly - a quantum system is constantly interacting with its environment.
Now, if microstate a is that the photon hits the left of my screen, then my screen (to set off a bomb) has to carry, say, a current along a wire all the way to London (kind of an inappropriate example, btw) to trigger a detonator to blow a primary charge to set off the bombs which will cause public mayhem leading to newspaper headlines in France read by someone who would have been killed by a bomb in Paris set off by primary charges detonated by a current carried from the right side of my screen triggered by the photon landing there. And at every single point, every moving electron and vibrating ion in the wires, every photon emitted by the bomb, every sound wave, every person witnessing it, reporting it, writing it, talking about it, printing it, reading it, every change to memory... each one of these is interacting with its environment and, taking the extension all the way, the entire Universe. Since we do not know the wave function of every single particle in the universe we cannot say which probabilities may be interfered out or docehered up or down. There could be a countless number of reasons that the photon hit the left of my screen, none of which have anything to do with the wave function of the photon or any component of my screen. (You've probably spotted I've changed my decoherence position to be a maxture of interference and decoherence.)

I admit, I don't know enough about it. That might all be gibberish, but I guess the rough concepts mean more to me than disembodied consciousness and multiple worlds. At the very least, you have got me fascinated in the possibility that interference and decoherence may be responsible for resolving the measurement problem, even if you think it's barmy. I will learn more, and maybe we can talk about it again.

I'm affraid that that's an error in Wikipedia, then. The only thing that phase coherence loss induces is the loss of observable interference effects. In fact, decoherence is based upon hopelessly STRONG entanglement!

Having thought about it myself, I'm inclined to agree.

(following up ; the post was too long to be submitted, so this is its second part).

Yeah, my bad. You can't call me apathetic or lazy, at least.

Because non-self aware systems can behave as if they are, and try to solve the problems associated with it.

No - if there's no link between consciousness and behaviour (and I couldn't agree less) then a non-conscious system cannot 'behave' as if it is conscious because any such behaviour would NOT be unique of consciousness and so the system is not behaving as if it has it.

BTW, humans are not known to be conscious. There's only one human I know it is, that's myself. I inductively ASSUME that others are.

And why not? you inductively assume that: 'quantum systems are in superposition' + 'we don't observe superposition' -> 'consciousness is a non-physical phenomenon that does not interact physically'.

But there can be evolutionary reasons to educate species to (sparely) kill themselves, if that inhances the survival chances of near kin (hence, increasing the genetic transfer to the next generation).

Do you know of any? If a species is aware of it's own existence so that it can end it, that's a good candidate for a non-human conscious species. If it is not aware, then it is not suicide. For example:

Lemmings kill themselves too, no ?

... is not an example of suicide, as the action (running towards the sea) that kills them ACCIDENTALLY is motivated by other... intentions.

Well, if it is one and the same consciousness, you should be aware of BOTH situations, which obviously is not the case, is it ?

No - your consciousness in the 'dead cat' branch will only experience 'dead cat', and likewise for 'live cat'. And once in that branch, it will experience that cat as dead forever more, since only dead cat states evolve from that state. However, live cat states are evolving in other branches, and your consciousness in those branches is experiencing live cat. Just treat consciousness as a physical system and repeat the examples you've already written - it should work!

Let me tell you what I really think: I don't really believe that "consciousness bull" myself.

You had me going. That was my main concern.

Now, from the very fact that human bodies end up entangled in wildly different macroscopic states (really, you don't believe me, but it is an elementary property, first noticed by Schroedinger, and of which I don't know how to escape - if someone knows, please tell me), clearly I don't consciously observe that.

You misunderstand me. Human bodies entangled in different states is not what alarms me. I'm not arguing with that.

From there follows almost immediately the dissociation of consciousness from the body as a degree of freedom (maybe not as a STATE). Photodetectors can happily be in superposition, because I'm not aware of them NOT being aware of their world in a superposition. It is just that *I* don't see that.

A recap on my two mutually exclusive thoughts on that:
- what you measure is not the quantum state of a system but a macroscopic observable property of it which is not necessarily in superposition;
- in MWI terms, if your consciousness IS entangled with the state of a system, the consciousness attached to one term will only observe that one term, while all other consciousnessesesses observe whatever one term they are attached to. Your consciousness IS in superposition, and each branch will evolve in such a way that a state of your consciousness will always evolve distinctly from the others. In other words, the are uncountable other Patrick's consciousnessesesses observing live cat, dead cat, happy cat, sour cat, tabby cat, ginger cat, cool cat, Cat from Red Dwarf, and so on. I don't subscribe to that, by the way, I'm just stating a possible MWI-friendly idea that does not involve disembodied consciousnessessesesses (aaarrrgggghhhh!), to show that they (may be) are unnecessary as a postulate.

If the man is inside, however, and remains there for years, he might start to really LEARN Chinese, with all the examples he has been treating, and utter finally his message with a few grammatical errors "my god this is boring, let me out!"

But only if he knows why he is in there and how he can get out and can predict outcomes he has no prior experience of - good consciousness-driven behaviour.

That's not necessarily the case. If you'd study in detail the brain and nerve reactions, you should in principle be able to completely predict the fact that if the picture of the hamburger is viewed long enough, this triggers neurons where associations are made with previous experience of eating such a hamburger, and which finally trigger the decision to go to McDonalds, eating one. This is then exactly the same as the computer scientist telling you: don't worry, your computer doesn't feel pain, look, here, is the line in the code where it says: "on space-bar event do" and look, here is the instruction to print on the screen "aauww, you hurt me".
So if you have a physical-chemical explanation of the behaviour of the human body (namely when seeing a picture of a hamburger, you go out to McDonalds eating one in this particular case), there is no "need" for consciousness to take that decision: the physics of the brain determine that. You simply passively EXPERIENCE that, and you THINK you made that decision, but you just experienced your brain physically acting in such a way that the decision was made. A physical reaction (which could be analysed by neurologists, in the same way as computer scientists could read the code of a computer) following the stimuli (and maybe including random generators) made your body buy a hamburger, and your consciousness was just given the feeling it decided it all by itself :-) You know, like with elections.

You completely misunderstood this. The 'buying the burger' IS triggered non-consciously - that's why it's called subliminal imagery! The conscious part is overriding this because you know WHY you want a burger, and so don't buy one. I don't think this can be explained without consciousness, and so consciousness can effect physical systems, such as his stomach. And arguing that this is all a physical thing is preaching to the converted - I believe ALL conscious thought is physically manifested.

How do we find out if computers are conscious or not ?

First, determine a definition of consciousness. Then devise a test that, of two similar systems, one conscious, one non-conscious, only one will pass. Give the computer the test.

That's not true, thinking about consciousness is as old as the world :-)

It's totally true. If consciousness cannot effect physical systems, then non-conscious humans could have evolved and the universe would be exactly the same... except for all them darn books, articles and forum-posts about consciousness, because in that universe there wouldn't be any to write about.

But I think you give "too much structure" to consciousness ! If you just call consciousness "what I observe" that will do, no ? It doesn't have to be a blue angle with yellow little wings !

On the other hand, you give it so little structure it becomes pretty meaningless. "what I observe" involves sensory input - that's physically explained. Left that vague, that might mean 'light projected onto the back of my retina'. It's no good, then, to define this phenomenon you think cannot be explained physically. In fact, saying something cannot be explained physically at all and yet not knowing what it is that cannot be physically explained is really a dead end. Also, when it comes to making this theory widely acceptable, there will be a burden of proof on you or other champions of it, since you are introducing entirely new concepts. I feel that such a non-definition of this phenomenon will be too encumbered by that burden to go anywhere.


The 'nature of consciousness' debate will probably go round and round, since you are firmly of the belief that it is non-physical and so cannot be detected while I am firmly of the belief it is physical and so may be detected, and neither of us have the support of empirical, hard evidence. Don't feel obliged to respond to every point (well, you're in no way obliged to respond to ANY of them, but... you know what I mean), but I'd be interested in being put right on the inteference/decoherence issue, so any input on that above all other points would be most appreciated.

 

[vanesch]

That's not what I said, and certainly not what I meant. I am the measuring system, and the thing I'm measuring is the screen - NOT the electrons. The screen has a macroscopic property of an interference pattern - the microscopic properties of where each individual electron landed are not of interest to me. This is as valid a macroscopic property as 'dead cat'/'live cat'.

Let's stick to this example:
apparently, you're now not considering a single impact as a measurement result, but only the accumulation of impacts, giving rise to an interference pattern. So let us consider a photographic film, which is exposed for a long time, and then develloped, as the measurement of interference.

Well, during the illumination, of course, the film gets in more and more superposed states, because it records individual impacts:

at t1, the film is in state:
|hit at x=0.5 and hit at x = 0.6> + |hit at x=0.23> + |hit at x=0.99 and hit at x=0.24 and hit at x=0.23>

at t2, more states occur in its state and so on... BUT...

of course the film has finite resolution (a finite number of AgBr xtals) and more and more terms start to MERGE AGAIN:

the term |hit at x=0.5 and hit at 0.6> followed by a hit at x=0.7 is indistinguishable (the same quantum state) as |hit at x=0.7> followed by a hit at 0.5 and at 0.6.

So as a more and more deterministic interference pattern builds up on the film, its many superposed states evolve back into only a few states, which all display about the same interference pattern: the difference is not macroscopic anymore.
If at that moment, you devellop the film, and you look at it, you have measured a certain state |interference level = 0.7>.
Indeed, your measurement basis now consists of states |interference level 0> (no interference), |interference level 0.1> ... and |interference level 1>.

You measured (almost) an eigenstate of the system, so you got a result for sure, and you don't "branch".

Good - this was my point. That is, the screen itself, as a quantum system, is still in a superposition of all the possible electron positions and momenta (which will effect the system as long the as momentum is transferred), but it's macroscopic property of 'interference pattern'/'no interference pattern' is NOT in superposition. So we can have quantum systems without macroscopic properties in superposition.

Of course. As long as the sub-hilbert space of microstates are eigenstates of the operator corresponding to the measurement of the macroscopic property, there's no problem with that, as I tried to illustrate.

Like I said, we got our wires crossed because this is not what I meant. I've clearly given the impression I'm even dumber than I am. \o/

Right, so I adapted the example. But it is not a very interesting one (or I'm missing your point) because we now measure a macroscopic observable of which our microsystem is in an eigenstate...

I understand the linearity argument. That's kind of my what my point was about earlier. I don't think microscopic states lead to macroscopic states for sure, because a macroscopic system is so much more complex than, say, a fundamental particle.

I don't understand that argument. If QM is postulated to be applicable to macroscopic systems, they have a (very complicated) Hilbert space, right ? And macroscopic properties are just observables over that Hilbert space which are vastly degenerate (many microstates correspond to the same macroscopic value of the property). That means that you can split up that very complex Hilbert space into a simple set of (orthogonal) eigenspaces, each corresponding to one single macroscopic value. So now we've put all that complexity in our pocket, and we work with these orthogonal eigenspaces of the macroscopic observable. The linearity argument works there, so we should end up into a superposition of two vectors, belonging to two different eigenspaces, which have two different macroscopic values. How can you avoid that ?

In one article of decoherence I read (which holds a completely different view to yours) it stated that:
1. decoherence evolves incredibly rapidly, such that the entire Universe's state is resolved quickly;
2. decoherence occurs constantly - a quantum system is constantly interacting with its environment.

Yes, I fully agree with that. I don't agree with the statement that decoherence makes a term disappear in the superposition, though. It quickly entangles both terms with such a complexity of other systems that UNDOING that entanglement (which is necessary in order to have a superposition only appearing in a Hilbert space of a controlled system, so that we can turn the basis there, and look at interference terms) is hopelessly impossible FAPP.

Now, if microstate a is that the photon hits the left of my screen, then my screen (to set off a bomb) has to carry, say, a current along a wire all the way to London (kind of an inappropriate example, btw) to trigger a detonator to blow a primary charge to set off the bombs which will cause public mayhem leading to newspaper headlines in France read by someone who would have been killed by a bomb in Paris set off by primary charges detonated by a current carried from the right side of my screen triggered by the photon landing there. And at every single point, every moving electron and vibrating ion in the wires, every photon emitted by the bomb, every sound wave, every person witnessing it, reporting it, writing it, talking about it, printing it, reading it, every change to memory... each one of these is interacting with its environment and, taking the extension all the way, the entire Universe. Since we do not know the wave function of every single particle in the universe we cannot say which probabilities may be interfered out or docehered up or down. There could be a countless number of reasons that the photon hit the left of my screen, none of which have anything to do with the wave function of the photon or any component of my screen. (You've probably spotted I've changed my decoherence position to be a maxture of interference and decoherence.)

What you just described correctly, is the huge number of interactions which lead many many systems to get entangled with our initial superposition, in a quite uncontrollable way - as you state. It is this which makes that interference effects, which depend on the INPRODUCT of both terms in the superposition, are bound to be 0, hence suppressing possible interference.

If you have 3 systems:

|U>|V>(|a> + |b>) you can hope to see something, when you limit your attention to the third system, by working in the basis |c> = |a> + |b> and |d> = |a> - |b>, for which you can set up a measurement. Looking at the projection on |c>, you find <a|a> + <b|b> + 2 Re(<a|b>) and it is the last term which gives you the interference.
However, after entanglement, you have
|u>|v>|a> + |w>|x>|b>. If you now do your experiment again, only limited to system 3 (because you cannot control systems 1 and 2), then the projection on |c> gives you: <a|a> + <b|b> + 2 Re(<a|b><u|w><v|x>)
And although <a|b> is still there, you now multiply it with <u|w> <v|x> ; given the complexity of these extra states, chances are that they are essentially orthogonal, whatever happens to them (time evolution). So you effectively suppress every kind of interference ; due to entanglement with "large and uncontrolled systems".

At the very least, you have got me fascinated in the possibility that interference and decoherence may be responsible for resolving the measurement problem, even if you think it's barmy. I will learn more, and maybe we can talk about it again.

Try to get a hold of the book "decoherence and the appearance of a classical world" by Joos and Zeh. It gives you a correct description of what decoherence does, and doesn't solve.

No - your consciousness in the 'dead cat' branch will only experience 'dead cat', and likewise for 'live cat'. And once in that branch, it will experience that cat as dead forever more, since only dead cat states evolve from that state. However, live cat states are evolving in other branches, and your consciousness in those branches is experiencing live cat. Just treat consciousness as a physical system and repeat the examples you've already written - it should work!

Well that's very good...
Only: you now seem to imply that there's a DIFFERENT consciousness in each branch ! That the "consciousness" in one term doesn't know anything about - face it, another - consciousness in another branch. GREAT. That's also what I'm saying: a conscious experience can apparently only experience ONE branch. Tell me why it does so, if you do not postulate that. Why does a consciousness experience only ONE branch ? Why not the entire wavefunction ?

And there is another problem with that (which MWI-ers still try to solve): if you associate a consciousness to each branch, MOST CONSCIOUSNESS EXPERIENCE DOES NOT EXPERIENCE THE BORN RULE ! Indeed, if there is a very small probability (according to Born) for something to happen, this will quantum mechanically result in:
0.9999 |a> + 0.001 |b> (or something of the kind).
But if |a> happily experiences that, and |b> too, and you do that experiment 1000 times, then, most terms will have about an equal number of |a> cases as |b> cases. So your average consciousness will experience a 50-50 result, and not a 10000 vs 1 result. They will not observe the Born rule.

You misunderstand me. Human bodies entangled in different states is not what alarms me. I'm not arguing with that.

Great. So why do we not experience that entanglement then ? Why doesn't it hurt when they hit our foot with a hammer in the other branch ?

A recap on my two mutually exclusive thoughts on that:
- what you measure is not the quantum state of a system but a macroscopic observable property of it which is not necessarily in superposition;
- in MWI terms, if your consciousness IS entangled with the state of a system, the consciousness attached to one term will only observe that one term, while all other consciousnessesesses observe whatever one term they are attached to. Your consciousness IS in superposition, and each branch will evolve in such a way that a state of your consciousness will always evolve distinctly from the others. In other words, the are uncountable other Patrick's consciousnessesesses observing live cat, dead cat, happy cat, sour cat, tabby cat, ginger cat, cool cat, Cat from Red Dwarf, and so on. I don't subscribe to that, by the way, I'm just stating a possible MWI-friendly idea that does not involve disembodied consciousnessessesesses

Well, you disembodied them already, because it is THE SAME BODY that is in different states in the superposition, but they are apparently DIFFERENT consciousnesses that experience those states. Good. I also went along that path. And now, the next problem is: where does the Born rule come from ? Most of these consciousnesses are in branches where experiments did wildly crazy things which are not in agreement with the Born rule.

cheers,
Patrick.

 

[vanesch]

You completely misunderstood this. The 'buying the burger' IS triggered non-consciously - that's why it's called subliminal imagery! The conscious part is overriding this because you know WHY you want a burger, and so don't buy one.

Ah, indeed, I misunderstood you. Your post was subliminal: I thought I had to eat a burger in every case :-)
Ok, then, in the second case, you can probably also find a physical explanation, that now more processing power is put into the system, associations with "publicity" "spending money" "getting fat" are triggered, and the neuron that has to decide whether to go to McDonalds or not gets now more "no" inputs that "yes" inputs, hence you don't do it. This processing is experienced as conscious thought, and you have the experience that you took the decision not to go to McDonalds.

First, determine a definition of consciousness. Then devise a test that, of two similar systems, one conscious, one non-conscious, only one will pass. Give the computer the test.

But the point is, of course, that in order to have your test qualified, you NEED already systems of which you KNOW that they are conscious, and that they aren't! And, apart from yourself, you don't really know that other systems are conscious. For instance, I can be of the opinion (showing my card of the KKK) that only white males (I love it to be politically not correct) are conscious. So I devise a test that looks at the reflectivity of the skin, the sex organs, have the subject lift 50 kg, make it run 500 meters, recite the 5 first chapters of Genesis and tell me 7 new racist jokes which make me laugh.
Most of my buddies, of which I think they are conscious, brilliantly pass the test. My tape recorder, my wife, the black people on the other side of the town, most members of the Democrat party... fail the test. I'm satisfied.
Until the day that I receive a white robot from a Japanese company with a kind of tube between the legs that can do exactly that. Hmmm, I say, technology is great, they make conscious robots now...

It's totally true. If consciousness cannot effect physical systems, then non-conscious humans could have evolved and the universe would be exactly the same... except for all them darn books, articles and forum-posts about consciousness, because in that universe there wouldn't be any to write about.

This is behaviour like any other ! What do you think happens in your brain when you write stuff about consciousness ? A good enough neurologist should be able to explain that too from the physical model of your brain. It's maybe a bug in our brains that has some evolutionary advantage, like thinking we should behave ethically and as such limit damage to our own species.

On the other hand, you give it so little structure it becomes pretty meaningless. "what I observe" involves sensory input - that's physically explained.

It doesn't necessarily involve sensory input. Think about what happens to people who had an accident, and who lost all of their senses: they don't hear, they don't see, they don't feel,... But if the doctor sees that their body is still functioning, and their brain is active, are they not conscious anymore ?
Consciousness cannot have any structure (physical degrees of freedom) because then it belongs to the body!

In fact, saying something cannot be explained physically at all and yet not knowing what it is that cannot be physically explained is really a dead end.

Yes, that's why it is - if I'm not mistaking - called the Hard Problem of consciousness

Also, when it comes to making this theory widely acceptable, there will be a burden of proof on you or other champions of it, since you are introducing entirely new concepts. I feel that such a non-definition of this phenomenon will be too encumbered by that burden to go anywhere.

I agree with you. It is not a theory, of course. It is an attempt to understand quantum theory. Nobody solved the riddle in principle of the interpretation of quantum theory under the conditions that I cited:

1) QM is correct and universal
2) the state vector describes the ontological world out there.

The 'nature of consciousness' debate will probably go round and round, since you are firmly of the belief that it is non-physical and so cannot be detected while I am firmly of the belief it is physical and so may be detected, and neither of us have the support of empirical, hard evidence.

I don't say it is non-physical in the sense of super-natural. I say it is non-physical in the sense that there is no Hilbert space associated with it - that it is not a degree of freedom that corresponds to a Hilbert space. Because in QM, all that is "physical" has a Hilbert space associated with it.
I say that it is something else, that is associated to one single term in the wavefunction of the universe, and to certain systems.

The simplest solution would be to dump all this consciousness stuff, and to say that the wavefunction, at a certain point, does collapse. That means then that unitary evolution is NOT strictly true. But, you will have to face it, there is not the slightest hint that this might be the case. And it gives hugely complicated problems with locality and causality, which ARE nicely solved in fully unitary QM.

cheers,
Patrick.

 

[Smurf]

so, which came first; The observer or the existee? (DUM DUM DUMM!)

 

[reilly]

What should I say to people who claim that QM supports the view that the universe couldn't exist without humans or some kind of "observer"?
......

First, tell them that QM makes no such claim whatsoever. What QM, and more generally science, do claim as important and required are empirical observations. If there are no observers, then who cares, why bother? The Occam prespective is -- don't sweat it, our universe has observers. Next?
There's no possible way to know about an observerless universe -- it's less than clear to me that the idea makes any sense at all.

The idea that humans are needed to account for or explain the universe is silly, absurd, ego driven nonesense. How could you ever know? How do you falsify such a claim -- theory will never be good enough to do so -- maybe once in 1,000,000 tries theory might do the trick.


............
Really?? I see where this comes from, but what is the truth? I'm not talking about getting into an ontological argument- just what QM really has to say on the subject. I see this so often and want to ask them how they happen to know what QM says, but I would rather know what you nice people think QM says so I can pass it along.
Many thanks.

.......
QM has nothing to say about the matter at all. Nor does science. We deal which what we perceive to "be out there." We are doing a pretty good job with the here-and-now of science -- computers, plastic Barbies, lasers, hybrid cars, microsurgery, Twinkies and 7/11 fried chicken with shelf lives of quite a few years -- maybe centuries. When we get past the day-to-day, it gets tougher, we are still just dealing with the basics in astronomy -- but the amount of observational data is sky rocketing, if you will pardon the phrase. It's fun to speculate on the big, big questions. But the plain fact is we don't really know much about the distant universe -- dark matter and inflation, 11 dimensional strings or whatever -- who knows. It's still early in the game of Universe. We really don't know enough to go to Universe for Who? (I'm often reminded of King Canute when considering Who's U?)

regards,
Reilly Atkinson

 

[El Hombre Invisible]

So as a more and more deterministic interference pattern builds up on the film, its many superposed states evolve back into only a few states, which all display about the same interference pattern: the difference is not macroscopic anymore.
If at that moment, you devellop the film, and you look at it, you have measured a certain state |interference level = 0.7>.
Indeed, your measurement basis now consists of states |interference level 0> (no interference), |interference level 0.1> ... and |interference level 1>.
You measured (almost) an eigenstate of the system, so you got a result for sure, and you don't "branch".
Of course. As long as the sub-hilbert space of microstates are eigenstates of the operator corresponding to the measurement of the macroscopic property, there's no problem with that, as I tried to illustrate.

I'm reading your argument in some weird polarized form, or maybe a superposition of two states. You seem to agree "the difference is not macroscopic anymore", which was my point, then say that we measure one eigenvalue out of 11 (0, 0.1.. 1). Not that the level of interference was a property I was interested in anyway. Are you saying they all have the same macroscopic properties, and agree therefore that a superposition of all possible states does not lead per se to a superposition of macroscopic observables? Because that's all my point was: that a quantum system may be in superposition, but the observed macroscopic property is evolved the same in each state. If you accept that it CAN happen, then the question is not why do we not observe quantum systems in superposition, but why do we not observe superpositions of macroscopic observables. The answer can then conceivably be: because they don't exist, and we're not stepping outside of QM. If these macroscopic properties are evolved in the same way as any state, then the cause should be moved from the realm of the supernatural to the workings of QM itself.

Right, so I adapted the example. But it is not a very interesting one (or I'm missing your point) because we now measure a macroscopic observable of which our microsystem is in an eigenstate...

Well, it's interesting to me. I'm new to all this.

I don't understand that argument. If QM is postulated to be applicable to macroscopic systems, they have a (very complicated) Hilbert space, right ? And macroscopic properties are just observables over that Hilbert space which are vastly degenerate (many microstates correspond to the same macroscopic value of the property). That means that you can split up that very complex Hilbert space into a simple set of (orthogonal) eigenspaces, each corresponding to one single macroscopic value. So now we've put all that complexity in our pocket, and we work with these orthogonal eigenspaces of the macroscopic observable. The linearity argument works there, so we should end up into a superposition of two vectors, belonging to two different eigenspaces, which have two different macroscopic values. How can you avoid that ?

I don't follow you. "you can split up that very complex Hilbert space into a simple set of (orthogonal) eigenspaces, each corresponding to one single macroscopic value"... "which have two different macroscopic values". Where did I get lost? A brick is a quantum system. All of the possible states it may be in are in superposition and we see 'brick'. If we split all those possible states apart into countless un-superpositioned descriptions, each one of those still say 'brick'. I mean, I don't know - it's just a thought. I just want to make a crowbar separation between 'superposition of states' and 'superposition of macroscopic observables' because I think this leap is probably incorrect.

|U>|V>(|a> + |b>) you can hope to see something, when you limit your attention to the third system, by working in the basis |c> = |a> + |b> and |d> = |a> - |b>, for which you can set up a measurement. Looking at the projection on |c>, you find <a|a> + <b|b> + 2 Re(<a|b>) and it is the last term which gives you the interference.
However, after entanglement, you have
|u>|v>|a> + |w>|x>|b>. If you now do your experiment again, only limited to system 3 (because you cannot control systems 1 and 2), then the projection on |c> gives you: <a|a> + <b|b> + 2 Re(<a|b><u|w><v|x>)
And although <a|b> is still there, you now multiply it with <u|w> <v|x> ; given the complexity of these extra states, chances are that they are essentially orthogonal, whatever happens to them (time evolution). So you effectively suppress every kind of interference ; due to entanglement with "large and uncontrolled systems".

If I understand you correctly: |a> is the result of the first system, |b> is the result of the second system, and |c> and |d> are the possible results of the third system. Only the third system is observed since the others are too complex. While we can guess that 2 Re(<a|b><u|w><v|x>) can be ignored, we're still left with <a|a> + <b|b>: a superposition. Trouble is, if we're not looking at these systems, we don't know what the probabilities of these outcomes are, or what happened within them to arrive at these outcomes. If I tell you that |a> and |b> are also orthogonal, due to some mysterious inner workings and inter-relations of those systems, how do you know if I'm right or wrong? This is exactly my point: even in this over-simplified scenario, we don't know what's going on to give us |a> and |b>. We're happy to work this stuff out for an electron in a hydrogen atom, but we can't possibly calculate these probabilities of these states in anything as complex as a cat, or probably even a half-mirror.

Try to get a hold of the book "decoherence and the appearance of a classical world" by Joos and Zeh. It gives you a correct description of what decoherence does, and doesn't solve.

Will do. Thanks. What level is this book aimed at? Anyway, I agree with you - decoherence alone cannot be the answer.

Only: you now seem to imply that there's a DIFFERENT consciousness in each branch ! That the "consciousness" in one term doesn't know anything about - face it, another - consciousness in another branch. GREAT. That's also what I'm saying: a conscious experience can apparently only experience ONE branch.

The major difference being that I'm suggesting consciousness is entangled to the rest of the wave function, rather than not being entangled at all and arbitrarily being assigned to one term of it.

Why does a consciousness experience only ONE branch ? Why not the entire wavefunction ?

A term of the 'consciousness' system can only experience the branch that it evolved in. It makes as much sense to ask why it cannot observe other terms, ones it is not entangled with, as it does to ask why 'live cat' cannot evolve into 'dead cat'.

And there is another problem with that (which MWI-ers still try to solve): if you associate a consciousness to each branch, MOST CONSCIOUSNESS EXPERIENCE DOES NOT EXPERIENCE THE BORN RULE ! Indeed, if there is a very small probability (according to Born) for something to happen, this will quantum mechanically result in:
0.9999 |a> + 0.001 |b> (or something of the kind).
But if |a> happily experiences that, and |b> too, and you do that experiment 1000 times, then, most terms will have about an equal number of |a> cases as |b> cases. So your average consciousness will experience a 50-50 result, and not a 10000 vs 1 result. They will not observe the Born rule.

But you have a much higher probability of being the consciousness in the 0.9999 branches than in the 0.001 branch, so you wouldn't know. Since you can't phone up your other consciousnesses and ask them which results they experienced, you'll never know if you have one that experiences an odd proportion of results. One of the descriptions of MWI I read said that the number of branches from an event for one outcome would reflect the probability of that outcome. For instance, there might be 9999 'a' branches and only one 'b' branch after the first experiment. That would mean 9999 consciousnesses experience 'a' while only one experiences 'b'. You are far more likely to be in an 'a' branch than a 'b', but there will be one in 10000^1000 branches in which a consciousness experiences 1000 'b's.

Great. So why do we not experience that entanglement then ? Why doesn't it hurt when they hit our foot with a hammer in the other branch ?

Your consciousness in the other branch will feel it. Your other states of consciousness have not evolved from 'hammer foot' states so will not experience that.

Well, you disembodied them already, because it is THE SAME BODY that is in different states in the superposition, but they are apparently DIFFERENT consciousnesses that experience those states. Good. I also went along that path. And now, the next problem is: where does the Born rule come from ? Most of these consciousnesses are in branches where experiments did wildly crazy things which are not in agreement with the Born rule.

I said kind of like different consciousnesses. Different states of the same consciousness would be the proper way to put it. The state 'left arm up, right arm down' will have a consciousness state entangled with that that experiences just that. The state 'left arm down, right arm up' will have another state aware of that.

Ah, indeed, I misunderstood you. Your post was subliminal: I thought I had to eat a burger in every case :-)
Ok, then, in the second case, you can probably also find a physical explanation, that now more processing power is put into the system, associations with "publicity" "spending money" "getting fat" are triggered, and the neuron that has to decide whether to go to McDonalds or not gets now more "no" inputs that "yes" inputs, hence you don't do it. This processing is experienced as conscious thought, and you have the experience that you took the decision not to go to McDonalds.

Keep going along those lines - you'll end up leaving no aspect of consciousness physical unexplained.

But the point is, of course, that in order to have your test qualified, you NEED already systems of which you KNOW that they are conscious, and that they aren't! And, apart from yourself, you don't really know that other systems are conscious. For instance, I can be of the opinion (showing my card of the KKK) that only white males (I love it to be politically not correct) are conscious. So I devise a test that looks at the reflectivity of the skin, the sex organs, have the subject lift 50 kg, make it run 500 meters, recite the 5 first chapters of Genesis and tell me 7 new racist jokes which make me laugh.
Most of my buddies, of which I think they are conscious, brilliantly pass the test. My tape recorder, my wife, the black people on the other side of the town, most members of the Democrat party... fail the test. I'm satisfied.
Until the day that I receive a white robot from a Japanese company with a kind of tube between the legs that can do exactly that. Hmmm, I say, technology is great, they make conscious robots now...

And that's precisely what people are doing right now, devising tests such as the spot test, which very few animals (and only humans above around 18 months old) seem to pass. Does this mean these animals are conscious? Not necessarily, and so the process goes on, refining the methods and our understanding. Your example would be a BAD method, which the scientific community as a whole would quickly debunk, so to get people to accept your theory you would have to continue to refine it until the evidence you present is really good enough, or just abandon your white supremacist theories and move on.

This is behaviour like any other ! What do you think happens in your brain when you write stuff about consciousness ? A good enough neurologist should be able to explain that too from the physical model of your brain. It's maybe a bug in our brains that has some evolutionary advantage, like thinking we should behave ethically and as such limit damage to our own species.

Is it? How many non-conscious animals do you know that have written about consciousness? Maybe it's a bug, maybe it's a reflex, maybe it's all just an amazing coincidence that non-conscious behaviour CAN make you talk about consciousness, AND consciousness is something else, AND it only seems to be the conscious ones that have this behaviour. It's not very convincing though. You're having to make too much up to make this theory work.

It doesn't necessarily involve sensory input. Think about what happens to people who had an accident, and who lost all of their senses: they don't hear, they don't see, they don't feel,... But if the doctor sees that their body is still functioning, and their brain is active, are they not conscious anymore ?

No, they're 'unconscious' if they are not consciously aware of anything. Anyway, you missed my point. I was saying such a vague definition could be interpreted simply as that, in which case pretty much any animal with eyes and any AI computer with a camera could fit that definition. By the way, we've moved on a lot since the times of 'we have five sense'. We now know the human body has a huge number of senses, most of which are not externally stimulated.

Consciousness cannot have any structure (physical degrees of freedom) because then it belongs to the body!

Not really an argument if you're of the opinion that it DOES belong to the body - that there's no difference between body and mind.

I don't say it is non-physical in the sense of super-natural. I say it is non-physical in the sense that there is no Hilbert space associated with it - that it is not a degree of freedom that corresponds to a Hilbert space. Because in QM, all that is "physical" has a Hilbert space associated with it.

And physics is the study of natural phenomena. Since all known physical phenomena are physical, the notion of a non-physical phenomenon is super-natural.

The simplest solution would be to dump all this consciousness stuff, and to say that the wavefunction, at a certain point, does collapse. That means then that unitary evolution is NOT strictly true. But, you will have to face it, there is not the slightest hint that this might be the case. And it gives hugely complicated problems with locality and causality, which ARE nicely solved in fully unitary QM.

From what little I know, I don't agree with that either, but I also don't see that just because I believe there is only one macroscopic observable then the system under observation is not in superposition. I just think making a cat analogous to a photon leads to misleading conclusions.