The witness-only channel and what physics allows

Write a scene in which a seer sees a room she has never entered, and the reader is left to decide whether the author has written a miracle or a cheat. The honest answer is that the physics already decided most of it decades ago, and it decided it in a way that is easy to misquote. Quantum entanglement creates correlations between distant events [1]. To date, all Bell tests have supported the theory of quantum physics, and not the hypothesis of local hidden variables [2]. And none of that lets anyone send a chosen message to anyone else, at any speed, ever. The seam between those two sentences is where this paper lives.

This paper states both facts precisely, gives the remote viewing and precognition literature its strongest hearing, gives the replications and the failed replications their strongest hearing too, and then answers one question, which is what physics allows when the channel in question is a witness-only channel, one that reads a correlation without transmitting or compelling anything.

What the theorem actually forbids

The no-communication theorem, also referred to as the no-signaling principle, is a no-go theorem in quantum information theory. It asserts that during the measurement of an entangled quantum state, it is impossible for one observer to transmit information to another observer, regardless of their spatial separation [1]. This conclusion preserves the principle of causality in quantum mechanics and ensures that information transfer does not violate special relativity by exceeding the speed of light [1].

The statement of the theorem is short, and the reason it holds is worth following in full, because almost every popular account of entanglement stops one line before it. The proof proceeds by defining how the total Hilbert space can be split into two parts, describing the subspaces accessible to two observers, Alice and Bob. The goal of the theorem is to prove that Bob cannot in any way distinguish the pre-measurement state from the post-measurement state. This is accomplished mathematically by comparing the trace of the two states, with the trace being taken over Alice’s subspace [1]. Since the trace is only over a subspace, it is technically called a partial trace [1].

Bqcfig3. Photo: P. Drmota, D. P. Nadlinger, D. Main, B. C. Nichol, E. M. Ainley, D. Leichtle, A. Mantri, E. Kashefi, and R. Srinivas et al. / Wikimedia Commons. Blind quantum computing: a SERVER and a CLIENT box joined by a classical link and a dotted quantum link. Source

The decisive sentence is the one that follows, and a reader who stops before it will misread the whole subject. Everything that Bob has access to, or could ever have access to, measure, or detect, is completely described by a partial trace over Alice’s subspace of the system. The fact that this trace never changes as Alice performs her measurements is the conclusion of the proof of the no-communication theorem [1]. Read that as a statement about what Bob can possibly learn. Not what Bob will learn, not what Bob may infer from a conversation afterwards, but what any measurement Bob could conceivably perform on his own wing can contain. Alice’s choice of apparatus, her choice of basis, her choice of how many times she measures: none of it appears there.

The question the theorem answers is worth quoting exactly, because it is the shape of the question a novel should ask. The question is: is there any action that Alice can perform on A that would be detectable by Bob making an observation of B [1]? The theorem replies ‘no’. The setup is equally plain. After a certain amount of time, the system is divided in two parts each of which contains some non-entangled states and half of the quantum entangled states, and the two parts become spatially distinct, sent to two distinct observers, Alice and Bob, who are free to perform quantum mechanical measurements on their portion of the total system [1].

There is one assumption in that setup that almost all of the argument’s critics forget, and it is the door through which everything interesting passes. An important assumption going into the theorem is that neither Alice nor Bob is allowed, in any way, to affect the preparation of the initial state [1]. The reason is trivial once stated. If Alice were allowed to take part in the preparation of the initial state, it would be trivially easy for her to encode a message into it [1]. The theorem does not require that the initial state be somehow ‘random’ or ‘balanced’ or ‘uniform’: indeed, a third party preparing the initial state could easily encode messages in it, received by Alice and Bob [1].

That sentence is the seam the whole question turns on, and it deserves to be read twice. A shared state may carry a message. A third party may put it there. Both observers may read it, and the reading may be correct, and the message may be information about something neither of them could otherwise have known. What neither observer may do is choose it, change it after the fact, or use the other wing’s readings to compose a new one. The theorem’s own summary is that, given some initial state, prepared in some way, there is no action that Alice can take that would be detectable by Bob [1].

What the theorem forbids, stated plainly, is narrower than the thing it is usually said to forbid. It is not possible to transmit classical bits of information by means of carefully prepared mixed or pure states, whether entangled or not [1]. A more careful statement is available, and it is the one a careful writer should use. The theorem is only a sufficient condition that states that if the Kraus matrices commute then there can be no communication through the quantum entangled states and that this is applicable to all communication. Being only a sufficient condition, there can be other reasons communication is not allowed [1].

And the phrase that has done the most damage to public understanding is here dismantled in one sentence. The no-communication theorem demonstrates that the failure of local causality does not imply that “spooky action at a distance,” a phrase often attributed to Einstein, can be used to communicate faster than light [1]. Nonlocal correlation and message are not the same thing, and the theorem is what separates them.

What Bell settled, and what it did not

The half of the story that a novel can use is real, and it should be told with its full strength. A Bell test is a real-world physics experiment designed to test the theory of quantum mechanics in relation to Albert Einstein’s concept of local realism [2]. It evaluates the implications of Bell’s theorem empirically [2]. As of 2015, all Bell tests have found that the hypothesis of local hidden variables is inconsistent with the way that physical systems behave [2].

That result is not a subtle one and it has not been withdrawn. To date, all Bell tests have supported the theory of quantum physics, and not the hypothesis of local hidden variables. These efforts to experimentally validate violations of the Bell inequalities resulted in John Clauser, Alain Aspect, and Anton Zeilinger being awarded the 2022 Nobel Prize in Physics [2]. Whatever a mystery novel wants its seer to be doing, it is not compatible with a universe in which local hidden variables explain the correlations.

Alain-Aspect-ForMemRS. Photo: Royal Society uploader / Wikimedia Commons. Alain Aspect, one of the three physicists awarded the 2022 Nobel Prize in Physics. Source

What fell is a specific proposal, and the way it fell is instructive. Einstein argued that the hidden variables should obey the condition of locality: whatever the hidden variables actually are, the behavior of the hidden variables for one particle should not be able to instantly affect the behavior of those for another particle far away [2]. Bell turned that intuition into a statistical limit. If the results of an experiment violate Bell’s inequality, local hidden variables can be ruled out as their cause [2]. What was ruled out was locality plus realism together. Nothing was ruled out about correlation as such.

The experiments are also harder to dismiss than the word intuition suggests. In 2015 an experiment was performed that is not subject to either the locality loophole or the detection loophole. An experiment free of the detection loophole is one where close to 100% of the successful measurement outcomes in one wing of the experiment are paired with a successful measurement in the other wing [2]. An experiment free of the locality loophole is one where for each separate measurement and in each wing of the experiment, a new setting is chosen and the measurement completed before signals could communicate the settings from one wing of the experiment to the other [2]. In practice, many present-day experiments employ the CHSH inequality [2].

What those inequalities actually constrain is worth stating, because it is narrower than the word locality suggests in popular use. All these inequalities, like the original devised by Bell, express the idea that assuming local realism places restrictions on the statistical results of experiments on sets of particles that have taken part in an interaction and then separated [2]. The restriction is on a joint distribution of numbers across many trials, not on any single event. Nothing in the result says that a particular pair of outcomes is impossible. It says that the pattern could not have been produced by a mechanism in which each wing carried its own hidden information from the moment the two wings separated. A character who is right about one thing is not contradicting Bell. A character who is right about a chosen thing, before the choice was made, is contradicting something far more basic.

The reason this is not the same as a working channel is stated, plainly, in the article on entanglement itself. Individual measurement outcomes cannot be controlled, so entanglement alone cannot be used to transmit information [3]. That is the whole of it. There is a correlation; the correlation is not under the writer’s control; therefore it cannot be composed into a message. Quantum nonlocality is described as the fact that measurements of a quantum system are correlated even when the systems are spacelike separated, and entangled states lead to correlations in the results of otherwise random measurements [4]. Correlation, yes. Composition, no.

The broader physical judgement is settled in the same way. The current scientific consensus is that faster-than-light communication is not possible, and to date it has not been achieved in any experiment [4]. A number of theories and phenomena related to faster-than-light communication have been proposed or studied, including tachyons, neutrinos, quantum nonlocality, wormholes, quantum entanglement and quantum tunneling [4]. The list is long and the consensus is unchanged.

The case for remote viewing, stated fairly

The other half of the story deserves the same care, because the case is not as foolish as its reputation and not as strong as its advocates say. Remote viewing is the practice of seeking impressions about a distant or unseen subject, purportedly sensing with the mind [5]. Precognition is the purported psychic phenomenon of seeing, or otherwise becoming directly aware of, events in the future [6]. Neither claim asks for a machine or for a force. Both ask for information that arrives without a channel, which is exactly the thing the theorem forbids when a channel is required and permits when it is not.

The programme is worth telling accurately, because the accurate version is more interesting than either the boast or the debunking. Physicists Russell Targ and Harold Puthoff, parapsychology researchers at Stanford Research Institute, are generally credited with coining the term “remote viewing” [5]. According to Targ, the term was first suggested by Ingo Swann in December 1971 during an experiment at the American Society for Psychical Research in New York City [5]. The subject then ran for two decades and more. The Stargate Project was a $20 million research program sponsored by the U.S. government that attempted to determine potential military applications of psychic phenomena [5]. The program ran from 1975 to 1995 and ended after evaluators concluded that remote viewers consistently failed to produce actionable intelligence information [5].

Russell Targ, physicist. Photo: photo was taken by subject’s wife (Patricia Targ) / Wikimedia Commons. Russell Targ, who coined the term remote viewing with Harold Puthoff. Source

That is a programme with funding, with methods, and with a published conclusion, and a writer should not pretend otherwise. What the conclusion was is just as much a part of it as the decades of work. The viewers’ advice in the Stargate Project was always so unclear and non-detailed that it has never been used in any intelligence operation [5]. A twenty-year programme that produced nothing anyone could act on, funded on the expectation that it might, is a fact about both the phenomenon and the institution that funded it.

The formal apparatus offered in defence is the most interesting part for a physicist’s novel, and it is older than its modern use. The two-state vector formalism is a description of quantum mechanics in terms of a causal relation in which the present is caused by quantum states of the past and of the future taken in combination [7]. It was first developed in 1955 by Satosi Watanabe, who proposed that information given by forwards evolving quantum states is not complete; rather, both forwards and backwards evolving quantum states are required to describe a quantum state [7]. Watanabe’s work was later rediscovered in 1964 by Aharonov, Bergmann and Lebowitz, who renamed it [7].

That pedigree is real, and it is the strongest card the parapsychology side holds. The formalism says that information given by forwards evolving quantum states is not complete; rather, both forwards and backwards evolving quantum states are required to describe a quantum state: a first state vector that evolves from the initial conditions towards the future, and a second state vector that evolves backwards in time from future boundary conditions [7]. A reader who wants a fictional physics in which knowing precedes the knowing can point at a time-symmetric account of measurement and be in good company. What the formalism does not do is license the transfer. It is an account of how a single measurement is described, not a device for sending anything anywhere.

The verdict on the whole field is stated without hedging in the source. Parapsychology is criticized as being a pseudoscience, and the majority of mainstream scientists reject it. It has been criticized for continuing investigation despite being unable to provide reproducible evidence for the existence of any psychic phenomena after more than a century of research [8]. The same article notes that research in the field rarely appears in mainstream scientific journals [8]. A century of work and no reproducible evidence is a fact, and a novel has to be written around that fact.

What the record shows

The failures are not a single scandal. They are a pattern, and the pattern is more useful to a writer than any one episode. Remote viewing experiments have historically lacked proper controls and repeatability [5]. That sentence is the shortest honest summary of the entire twentieth-century literature, and it does not require a theory of fraud to explain the record.

The first failure mode is the simplest and the oldest: information reaching the viewer by an ordinary route. One of the early experiments, lauded by proponents as having improved the methodology of remote viewing testing and raising future experimental standards, was criticized as leaking information to the participants by inadvertently leaving clues. Some later experiments had negative results when these clues were eliminated [5]. A scene in a novel in which the seer is right is worth a great deal less if the author has not shown how the seer could have been wrong.

The second is a straightforward negative result obtained by physicists rather than sceptics. In the late 1970s, the physicists John Taylor and Eduardo Balanovski tested the psychic Matthew Manning in remote viewing, and the results proved “completely unsuccessful” [5].

The third is the criterion the sceptical physicist actually applied. Hyman’s conclusion was that by themselves these experiments do not mean anything unless they can be replicated [9]. He then asked parapsychologists “What is the best evidence for psi?” they nearly universally pointed to the Ganzfeld experiment. Hyman wrote to Charles Honorton and was sent 600 pages of information [9]. The criterion is the one the field has never fully met.

The fourth and most instructive is what happened to the most famous modern claim. One of the nine experiments in Bem’s study was repeated by Ritchie, French and Wiseman, and their attempt to replicate was published and found no evidence of precognition [10]. Later a large-scale pre-registered replication in 2023, which included Bem and other proponents of extrasensory perception in its design, found no evidence for precognition [10]. The interesting part is what happened around the replication papers rather than to them. Several failed attempts by the authors to publish their replication attempt highlighted difficulties in publishing replications, attracting media attention over concerns of publication bias [10].

Funnel plot depicting asymmetry Sept 2015. Photo: Sardmeta1 / Wikimedia Commons. Asymmetric funnel plot: all study markers lie left of the ln ES = 0 centre line, the signature of publication bias. Source

The last is a sentence from the man himself, and a writer should know it exists before inventing something harsher. Asked about the difficulty of moving into a field where the data count for so much, Bem said he preferred other people to do it, and that if you looked at all his past experiments they were always rhetorical devices and he used data as a point of persuasion. A colleague saw this as coming remarkably close to an outright admission of p-hacking [10]. Whatever one makes of it, it is a description of how a body of results came to be believed.

The field-level verdict on precognition is the same as on remote viewing, and it is not close. There is no accepted scientific evidence that precognition is a real effect, and it is widely considered to be pseudoscience [6]. The claim that it violates the principle of causality is a separate and additional objection, and it is the one that connects directly to the physics in the first section.

Precognition also carries a difficulty that remote viewing does not, and the difficulty is structural rather than statistical. Precognitive dreams are the most widely reported occurrences of precognition. Usually, a dream or vision can only be identified as precognitive after the putative event has taken place [6]. The event is the test, and the test cannot be applied until the thing being tested has already happened. A hypothesis in that position cannot fail at the moment it is made, which is the definition of a claim that survives by its timing rather than by its evidence. A novel can use that structure knowingly, as a way of showing how a person becomes convinced, and cannot use it as a way of being right.

Why the quantum vocabulary does not travel

The most common failure in popular accounts is not dishonesty but a category error, and it is worth naming precisely because a writer can avoid it. The move usually goes: entanglement connects distant things; a seer connects distant things; therefore the seer may use entanglement. Every step after the first is unsupported. Individual measurement outcomes cannot be controlled, so entanglement alone cannot be used to transmit information [3].

The first step of the counterargument is a property of entanglement that has no counterpart anywhere in the psychology of a seer. In quantum physics, monogamy is the property of quantum entanglement that restricts entanglement from being freely shared between arbitrarily many parties [11]. In order for two qubits A and B to be maximally entangled, they must not be entangled with any third qubit C whatsoever [11]. If two parties are maximally entangled with each other, no third system can be entangled with either of them at all.

Navaho Dilution Refrigerator. Photo: OJB Quantum / Wikimedia Commons. Dilution refrigerator stages from 300 K down to 10 mK, labelled in Navajo. Source

Apply that to a proposed mind-to-mind channel. In order for two qubits A and B to be maximally entangled, they must not be entangled with any third qubit C whatsoever [11]. That applies to the apparatus, to the target’s own mind, and to the record the seer will later compare. A relationship that is total is not a relationship that can be exploited in parallel. Whatever a remote viewer is doing, whatever the coupling is, it is not the coupling of two maximally entangled systems, because no third system is left over to read the result.

The second step is about what the formalism actually delivers when it is used properly, and the example is the cleanest available. Quantum teleportation only transfers quantum information. The sender does not have to know the particular quantum state being transferred [12]. It is often read as the counterexample to no-communication, a case where something arrives at a distance. It is not, because classical information needs to be sent from sender to receiver, and because classical information needs to be sent, quantum teleportation cannot occur faster than the speed of light [12]. The exotic part travels; the boring part carries the message; and the message is what the receiver actually needs.

The third step returns to the sentence that ends the discussion of every attempt to use entanglement as a telephone. Individual measurement outcomes cannot be controlled, so entanglement alone cannot be used to transmit information [3]. A writer can state the rule as a rule of the world without naming it. Choose your own apparatus at a distance, and you have chosen nothing that your counterpart on the far side can see.

There is a fourth point, weaker than the first three but worth having. The distance and timing of the measurements can be chosen so as to make the interval between the two measurements spacelike, hence any causal effect connecting the events would have to travel faster than light [3]. It is not even possible to say which of the measurements came first [3]. A theory built on a seer who knows before the event, in a spacetime where the ordering of two events is a matter of which inertial frame you ask, is a theory with a load-bearing ambiguity, and the ambiguity is not resolvable by adding sincerity.

The seam

Knowing at a distance, signalling at a distance and causing at a distance are three different things, and the entire question turns on keeping them apart. The theorem is only a sufficient condition that states that if the Kraus matrices commute then there can be no communication through the quantum entangled states [1]. Knowing is what the theorem has nothing to forbid. A correlation exists; a shared record exists; a third party may even have written something into it. Both parties may later discover that their records agree in a way neither could have arranged. That is not communication, and the theorem does not touch it. Signalling is what the theorem forbids outright. It is the transmission of a chosen symbol, and no action on one side can be arranged to be detectable on the other.

Single-Photon-Counting-Module Excelitas. Photo: Geek3 / Wikimedia Commons. Excelitas single-photon counting module, the plug-in detector used on photon coincidence benches. Source

Causing is a third thing again, and it is the one a novel is most tempted to write. To cause an outcome at a distance is to choose it, and choosing is exactly the operation the trace argument rules out. A seer who cannot be argued with, cannot be tested, and cannot be proved wrong because the outcome only exists once it has been read, is not operating outside physics. They are operating outside the condition that made the subject checkable in the first place. That condition is this: individual measurement outcomes cannot be controlled, so entanglement alone cannot be used to transmit information [3].

So the answer to the question the owner actually asked is a qualified yes, and the qualification is the whole of the paper. Established physics does allow a witness-only relation: two parties can hold a shared state, can each observe their own wing, and can later compare records and find a correlation that neither could have chosen and neither could have altered. A third party preparing the initial state could easily encode messages in it, received by Alice and Bob [1]. That is a message neither party composed. It is a witness, not a command.

And established physics destroys remote viewing as ordinarily claimed, in one sentence. If the claim is that a viewer can receive specific, freely chosen information about a chosen target, then the claim is a signalling claim. The fact that this trace never changes as Alice performs her measurements is the conclusion of the proof of the no-communication theorem [1]. The distance does not matter. The entanglement does not matter. Monogamy takes care of the rest. In order for two qubits A and B to be maximally entangled, they must not be entangled with any third qubit C whatsoever [11].

The line, then, is this. Perception that reads a correlation already present is permitted, and no experiment can distinguish it from the alternative explanations. It is not possible to transmit classical bits of information by means of carefully prepared mixed or pure states, whether entangled or not [1]. Perception that carries a chosen message is therefore forbidden, and the prohibition is not a limitation of current technology but a property of the formalism. And perception that compels an outcome is the same operation with a different name, because choosing the measurement and choosing the outcome are the same act from the point of view of the observer who cannot control it.

The distinction has a concrete shape, and a writer can draw it as a sequence. Quantum teleportation only transfers quantum information, and the sender does not have to know the particular quantum state being transferred [12]. Two parties hold a shared state and each measures their own wing. The result of the measurement is sent across an ordinary channel, and the receiver cannot reconstruct anything until those classical bits arrive [12]. The two parties can then compare what they recorded, by entirely ordinary means [2]. That procedure uses entanglement for what entanglement is good for, and uses an ordinary channel for what only an ordinary channel can do. Every step of it is permitted, and the theorem is what makes the outcome surprising rather than what makes it impossible. Now change one thing. Let the first party, having learned what the second party is about to measure, choose her own setting in order to put a particular symbol into the record. That single change is the step that turns the description from a permitted procedure into a signalling scheme. Everything that the far wing has access to is completely described by a partial trace, and that trace never changes as the first party measures [1]. The theorem does not say the first procedure is impossible. It says the second one has no channel.

Connecting the Dots

The threads run like this. A Bell test is a real-world physics experiment designed to test the theory of quantum mechanics in relation to Albert Einstein’s concept of local realism [2]. The result has held through every test since, including one performed in 2015 that is not subject to either the locality loophole or the detection loophole [2]. The same body of physics produced a theorem about two observers holding one half of an entangled pair each. Given some initial state, prepared in some way, there is no action that Alice can take that would be detectable by Bob [1]. Both statements are settled, and they are not in tension with each other. They are in tension with the sentence that people usually write between them.

That middle sentence is the whole problem, and it has three words that do different work. Nonlocal describes the correlations and is true. Sending describes the transfer of a chosen symbol and is false. Forcing describes the selection of an outcome and is false in the same way for the same reason. Individual measurement outcomes cannot be controlled, so entanglement alone cannot be used to transmit information [3]. A paper that uses the first word and slips into the second has not found a loophole in physics. It has found the place where a reader stops checking.

Wikipedia and Wikidata language correlations. Photo: Thomas Shafee / Wikimedia Commons. Correlation matrix of Wikipedia and Wikidata metrics with scatter plots and trend lines. Source

The second thread is a documentary one, and it belongs in the same dossier because it is about the same seam. Twenty million dollars, and twenty years of it. The program ran from 1975 to 1995 and ended after evaluators concluded that remote viewers consistently failed to produce actionable intelligence information [5]. The same record contains two more findings. Remote viewing experiments have historically lacked proper controls and repeatability [5]. One of the early experiments was criticized as leaking information to the participants by inadvertently leaving clues. Some later experiments had negative results when these clues were eliminated [5].

The third thread is the one a novelist can actually use. The two-state vector formalism says the present is caused by quantum states of the past and of the future taken in combination [7]. That is a description of a measurement in which the future conditions participate in the present account. It is not a claim that the future can be edited, and the difference is the difference between an axiom and a device. A character may stand on the formalism and be interesting. She may not stand on it and be right.

The Other Side

The strongest version of the case goes like this. A programme that ran from 1975 to 1995 and ended after evaluators concluded that remote viewers consistently failed to produce actionable intelligence information [5]. A century of research in the field has produced no reproducible evidence for the existence of any psychic phenomena [8]. That is where this argument has to start. Small and contested is not the same as absent, and a physics that rules out signalling does not rule out a correlation that nobody has yet explained.

The strongest objection to that is not that the data are wrong but that the category is. It is not possible to transmit classical bits of information by means of carefully prepared mixed or pure states, whether entangled or not [1]. If the phenomenon is a correlation, it is a correlation like the ones the field already explains, and the burden is on the phenomenon to show it is not. If the phenomenon is a transmission, it is a transmission, and the burden is on it to explain how the trace on the far side could contain the choice.

There is a second objection that the historical record supplies. The field has been unable to provide reproducible evidence for the existence of any psychic phenomena after more than a century of research [8]. A century is a long time, and it is long enough that the ordinary explanations, chance, leakage, flexible analysis, and the pressure to publish something interesting, have all been tested and have all been sufficient. The hypothesis that survives is the one that asks for a new mechanism. That is not a reason to refuse the hypothesis. It is a reason to want more than an argument about what the mechanism would be.

Region of Hubble Ultra Deep Field (Digitized Sky Survey). Photo: Digitized Sky Survey (DSS), STScI/AURA, Palomar/Caltech, and UKSTU/AAO / Wikimedia Commons. Digitized Sky Survey star field of the Hubble Deep Field region (STScI/AURA, public domain). Source

What We Actually Know

Very nearly certain is the state of the question. The current scientific consensus is that faster-than-light communication is not possible, and to date it has not been achieved in any experiment [4]. Almost certain is that local hidden variables have been ruled out as an explanation of the correlations Bell tests measure, and that the 2022 Nobel Prize in Physics was awarded for the experimental validation of the violations [2]. Almost certain is that remote viewing experiments have historically lacked proper controls and repeatability, and that the programme which ran from 1975 to 1995 ended with its evaluators concluding that remote viewers consistently failed to produce actionable intelligence information [5].

Very likely, the phenomenon is real and the explanation is not the one being offered. Monogamy is a property of quantum entanglement, and in order for two qubits to be maximally entangled they must not be entangled with any third qubit whatsoever [11]. Whatever a viewer perceives, it is not that, and any theory that needs it to be that is asking for a physics nobody has.

Very unlikely is any form of the claim that a chosen message has been transmitted. The distance and timing of measurements can be chosen so as to make the interval between them spacelike, and it is not even possible to say which of the measurements came first [3]. A claim that depends on the ordering of two events that have no agreed order is not a claim that can be tested, whatever its merits.

Detector arrays of a Pegasus satellite. Photo: Unknown authorUnknown author / Wikimedia Commons. Pegasus satellite meteoroid-detector array under assembly in a Fairchild cleanroom (NASA, public domain). Source

What we can actually know is narrower than either camp wants. A witness-only relation is permitted. A third party preparing the initial state could easily encode messages in it, received by Alice and Bob [1]. Nobody composes that message. Signalling is forbidden. The fact that this trace never changes as Alice performs her measurements is the conclusion of the proof of the no-communication theorem [1]. And causing is not a third possibility. It is signalling with the requirement of a result attached, and the same proof covers it. Individual measurement outcomes cannot be controlled, so entanglement alone cannot be used to transmit information [3].

References

[1] No-communication theorem — https://en.wikipedia.org/wiki/No-communication_theorem (accessed 2026-10-06) [Reputable]

[2] Bell test — https://en.wikipedia.org/wiki/Bell_test (accessed 2026-10-06) [Reputable]

[3] Quantum entanglement — https://en.wikipedia.org/wiki/Quantum_entanglement (accessed 2026-10-06) [Reputable]

[4] Faster-than-light communication — https://en.wikipedia.org/wiki/Faster-than-light_communication (accessed 2026-10-06) [Reputable]

[5] Remote viewing — https://en.wikipedia.org/wiki/Remote_viewing (accessed 2026-10-06) [Reputable]

[6] Precognition — https://en.wikipedia.org/wiki/Precognition (accessed 2026-10-06) [Reputable]

[7] Two-state vector formalism — https://en.wikipedia.org/wiki/Two-state_vector_formalism (accessed 2026-10-06) [Reputable]

[8] Parapsychology — https://en.wikipedia.org/wiki/Parapsychology (accessed 2026-10-06) [Reputable]

[9] Ray Hyman — https://en.wikipedia.org/wiki/Ray_Hyman (accessed 2026-10-06) [Reputable]

[10] Daryl Bem — https://en.wikipedia.org/wiki/Daryl_Bem (accessed 2026-10-06) [Reputable]

[11] Monogamy of entanglement — https://en.wikipedia.org/wiki/Monogamy_of_entanglement (accessed 2026-10-06) [Reputable]

[12] Quantum teleportation — https://en.wikipedia.org/wiki/Quantum_teleportation (accessed 2026-10-06) [Reputable]