Why can't Spooky action send FTL information?

[amoses7178]

Why can't "Spooky action" send FTL information?

I understand that for two entangled particles, that if you observe one it will instantly affect the other. I can even understand how we can't make sense of what was observed without sending a "Hey, I got and 'up' what about you?"

But what I don't understand is why information itself can't be sent instantly using this method.

After all, what if you were holding an entangled particle in superposition and you watched it carefully without directly observing it (such as with Wigner's Friend a.k.a the "Quantum Mouse")? Wouldn't the very fact that the particle's wave function collapsed generally give you a clue that its partner had been observed?

And if that's true, what would prevent putting a sequence of entangled particles together so that the wave function collapse or lack there-of encoded a message?

-Alan

 

[mcah5]

My guess would be that it is impossible to "indirectly observe" the wavefunction without always collapsing the wavefunction. So the receiver would have no way of knowing whether the antiparticle had already been measured or not.

 

[vanesch]

And if that's true, what would prevent putting a sequence of entangled particles together so that the wave function collapse or lack there-of encoded a message?

The answer to this question (which is often posed) is given in my journal (to the left of this post). It comes down to the fact that (no matter what interpretation of QM you favor) the local reduced density matrix on one side (which determines all probabilities of all outcomes of measurements done purely on that side of the system: no "correlation" measurements, but just local measurements) is INVARIANT under a measurement on the other side or not. This means that locally the probabilities of everything you can measure, do not change, no matter what measurement is done on the other side. So locally, you cannot know what the OTHER side did to its system, and hence not use this as a communication channel.
The "spooky action at a distance" is only observable if you look at CORRELATIONS between measurements on both sides (then you cannot use the REDUCED density matrix anymore, but you need to use the FULL density matrix). But on one side, you cannot use these correlations, you can only use local measurements. And they don't change.

cheers,
Patrick.

 

[Antiphon]

And if that's true, what would prevent putting a sequence of entangled particles together so that the wave function collapse or lack there-of encoded a message?

In the past on this forum I've advanced a special relativity argument
against sending such information. The short version is this- in order
to send information though a channel like this, the sending side must
be a causitive agent in the "collapse". This implies that the sending
side make it's measurement "first" or "before" the receiving side. But
special relativity says that the relative order of events which are widely
separated depends on the reference frame, and it is truely and
physically meaningless to talk about the order of spatially separated events.

I can construct a reference frame with an observer near the "transmitter"
in which the "receiver" is seen to make their measurement first.

This is why the results are always correlated but neither side can be said
to have caused the collapse.

I never cease to mavel at how relativity and QM are so consistent even
in the most bizarre situations.

 

[amoses7178]

Is it not possible to have a particle in superposition and know *when* it collapses?

The idea is to make timed observations, the timing and sequence of which encodes “bits” of True/False which could be laid out like Morse Code.

Even if there is some uncertainty for exactly when an observation is made, so as long as we know one has collapsed and another has not, wouldn’t it still be possible to string together a pre-arranged sequence of these observations that contain “bits” of True/False?

 

[selfAdjoint]

Is it not possible to have a particle in superposition and know *when* it collapses?

The idea is to make timed observations, the timing and sequence of which encodes “bits” of True/False which could be laid out like Morse Code.

Even if there is some uncertainty for exactly when an observation is made, so as long as we know one has collapsed and another has not, wouldn’t it still be possible to string together a pre-arranged sequence of these observations that contain “bits” of True/False?

You're missing the point that the observer of the first collapse can't control which eigenstate it will collapse into. Therefore when the observer of the second collapse sees some eigenstate she can't conclude (being out of touch because spacelike to the first one), that her collapse came from a particular eigenstate at the other end. So she can't get a bit from that channel.

 

[Sherlock]

In the past on this forum I've advanced a special relativity argument against sending such information. The short version is this- in orderto send information though a channel like this, the sending side must be a causitive agent in the "collapse". This implies that the sending
side make it's measurement "first" or "before" the receiving side. But
special relativity says that the relative order of events which are widely
separated depends on the reference frame, and it is truely and
physically meaningless to talk about the order of spatially separated events.

I don't think this is why you can't send info superluminally or
instantaneously between A and B. In the lab frame all the
components have a fixed relationship to each other, and a
common clock is used. In this case you can say unambiguously,
without relativity, which detection of a pair occurred first.
But, this is irrelevant anyway.

The reason you can't send info superluminally or instantaneously
between A and B is because they aren't causally related to
each other.

 

[Antiphon]

I don't think this is why you can't send info superluminally or
instantaneously between A and B. In the lab frame all the
components have a fixed relationship to each other, and a
common clock is used. In this case you can say unambiguously,
without relativity, which detection of a pair occurred first.
But, this is irrelevant anyway.

The reason you can't send info superluminally or instantaneously
between A and B is because they aren't causally related to
each other.

Yes, I agree that in the lab frame you know which came first.
But the lab frame is not preferred in any way. I make the almost
philisophical argument that if you could transmit information
this way, it would lend special weight to those frames in which
the transmitter's measurements came first. SR denys that such
a preferred frame(s) exists.

Edit: While I agree that A and B are not causually related, they are
invariably correlated. It's this correlation which lures people into thinking
they can employ it as a means of communicating.

 

[Sherlock]

Yes, I agree that in the lab frame you know which came first.
But the lab frame is not preferred in any way. I make the almost
philisophical argument that if you could transmit information
this way, it would lend special weight to those frames in which
the transmitter's measurements came first. SR denys that such
a preferred frame(s) exists.

Edit: While I agree that A and B are not causually related, they are
invariably correlated. It's this correlation which lures people into thinking
they can employ it as a means of communicating.

A and B aren't correlated to each other. The joint measurement,
(A,B) is correlated to some joint variable, eg., the angular difference
between polarizer settings.