Quantum Mechanics does not “Deny Reality”: Contextuality is Realistic
Does reality need to be independent of perspective in order to be real?
Quantum mechanics is the subject of a lot of quantum woo and mysticism. A very common claim is that quantum mechanics somehow calls into question the very existence of an objective reality independent of the observer. There is even currently a popular article from a few years ago on this website stating this, titled “Why is no one taught the one concept in quantum physics which denies reality?” Yet, there is simply nothing about the theory that comes close to implying this at all.
Claiming that contextuality demonstrates that we do not perceive objective reality as it exists independent of our observations of it is an incredibly bizarre claim. Indeed, how we “look” at something changes what we “see,” but we knew this going back to Galilean relativity. Two observers in different frames of reference can see the same object with a different velocity. It has always been, since the very earliest days of science, true that you cannot specify the properties of a system without specifying the context under which they are being observed, the “point-of-view” so to speak.
Yet, this does not prove “reality does not exist independent of the observers.” Point-of-view is more like a coordinate system. It is like taring a scale. You choose an object as the basis of a reference frame from which you describe other objects in relation to it. The fact you have to specify a coordinate system, a point-of-view, in no way makes it dependent upon conscious observers. When you measure the velocity of an object, you are revealing the object’s properties as it objectively exists independently of the observer, but dependent upon the point-of-view of that observation.
Against Hidden Variable Contextuality
Consider the principle of complementarity. If you have two complementary variables, like position and momentum, you can only know one at a time. If you measure one of them, the other becomes unknown. For example, if you measure position, then the momentum becomes unknown.
Without hidden variables, the reason the momentum becomes unknown is because it genuinely just ceases to exist: the particle just no longer has a definite momentum. However, with hidden variables, the particle would indeed have a momentum, but you just wouldn’t know what it is.
There are multiple usages of the word “contextuality,” but in some usages, it refers to a hidden variable theory which posits that systems have well-defined properties at all times, but you just can’t know them. In this sense, I will refer to it as hidden variable contextuality. How would that work? Well, consider, for example, if when you measure the position, you perturb the momentum due to the observer effect. Since you disturb the system by measuring it, your measurement changes the value in a way that is unpredictable.
Indeed, you can replicate many behaviors of quantum mechanics by presuming systems have internal states that are disturbed by the act of measurements. See this paper and this paper. Hidden variable contextuality is not even a property of quantum mechanics but also of classical mechanics. It only becomes a property of quantum mechanics because it still seems to apply even when measurement results are distributed over vast distances, and thus, if you believe in hidden variables, you would have to posit non-local effects caused by measurement disturbances.
This case with a single particle being disturbed by measurement is a very simple case, I would recommend reading my other article here that discusses a more complex case with a triplet of particles. If you try to assign them all definite values simultaneously, you find it is impossible to do so without taking into account what is actually measured by the observer.
In other words, measurement results would not be merely revealing properties that are already there, but the act of measurement would also be disturbing those results, and thus the actual choice of measurement itself would have to be included in whatever hidden variable theory predetermines the outcome.
Yet, quantum mechanics is not a hidden variable theory. There is no presumption of measurement disturbance. There is no observer effect. This is a rather subtle point as the most obvious and intuitive way to explain the uncertainty principle is with the observer effect, yet this explanation falls apart due to Bell’s theorem, and thus quantum mechanics should be interpreted without this assumption.
If there is no observer effect, then you are indeed revealing the states of particles as they exist. I’ll have to dedicate the next section to explaining why the observer effect is a problematic explanation and why it cannot even be justified, as it is a very common mistake and a subtle point that is not at all obvious.
Against the Observer Effect
The term “observer effect” refers to a system being perturbed, and thus changing its state, based on the very act of trying to observe it. It exists in many fields of science, and not merely quantum mechanics. For example, in psychology, there is something known as the Hawthorne effect, whereby someone being studied changes their behavior as a result of being studied.
If, for example, a person does indeed change their behavior as a result of being studied, then you cannot extrapolate what you learn from studying them to conclude upon how they must behave independent of your observation. Indeed, if we believe that there is an observer effect in quantum mechanics, then, similarly, we cannot conclude from what we observe the properties of reality independent of our observation of it. This is where the claims that quantum mechanics is not compatible with realism originate from.
How do we even know if there is an observer effect? Take the example of the Hawthorne effect. How can you actually be certain your observation is changing the test subject’s behavior? The only way you could verify this is with a more subtle measurement. If you conducted two experiments where in one you observed the test subject in secret, with great subtlety, and in the other, you observed them openly without hiding it, and you found their behavior changed drastically, you could conclude from this that your less subtle measurement is altering their behavior.
Yet, in quantum mechanics, there is no such thing as a “more subtle” measurement. Energy comes in quantized packets, there is an absolute minimum of subtlety of measurements. If such an observer effect existed, then it would be impossible to ever demonstrate.
Maybe, you might think, there is an indirect way of proving there is an observer effect. Consider, for example, the statistical laws of independence and dependence. You can compare if two events are statistically independent or statistically dependent simply by verifying that, in a probabilistic experiment, whether they fit the first equality below. If they do not, they should fit the second equality, and thus the events would be statistically dependent.
Surely, if the outcome of an experiment is statistically dependent upon what is measured, then it clearly there is a measurement disturbance, yes? This is what, I believe, is the implicit thought process going on in people’s heads when they use something like the double-slit experiment as evidence in favor of the observer effect.
Consider a simpler example of the double-slit experiment: the Mach — Zehnder Interferometer. A photon hits a beam splitter, and then it is reflected off of two mirrors and rejoined back at a second beam splitter. In the diagram below, if the photon’s path between the two beam splitters is not measured, then there is a 100% probability of the photon showing up at detector 2 and a 0% probability of it showing up at detector 1 due to interference effects.
If one photon’s path is measured in between the two beam splitters (such as placing a measurement device where “sample” is in the diagram above), then the probability distribution actually changes: there would be a 50% chance of the photon showing up on detector and a 50% chance of it showing up on detector 2.
Consider an experiment where, randomly, 50% of the time, a measurement is made of the photon’s path between the two beam splitters. We can call this B. Each time, the outcome on detector 2 is also measured. We can call this A. P(B)=50% as we defined it such that it occurs 50% of the time. P(A)=75% as when B occurs, A occurs 100% of the time, but when B does not occur, A occurs 50% of the time. P(A∩B)=25% as this only occurs when there is both a measurement of the photon’s path between the two beam splitters, and it happens to show up on detector 2.
In this case, P(A)P(B)=37.5% and thus P(A∩B)≠P(A)P(B). Obviously, P(A|B)=50% as we already stated that the probability of detector 2 measuring a photon is 50% given that the photon’s path between the two beam splitters is measured. Hence, P(A∩B)/P(B)=25%/50% which is 50%, and thus it satisfies the conditional probabilities of P(A|B)=P(A∩B)/P(B).
Even though people do not vocalize this reasoning, I believe this is implicitly what is going on in people’s heads when they state that experiments like the double-slit experiment prove there is an observer effect. There seems to be a statistical dependence upon measuring the midway path of the photon, such that the statistical results are different if you do not measure it.
Yet, this argument is actually rather flawed and cannot determine whether there is an observer effect. Consider a simple experiment with a moving train traveling at 100 km/h. If you get in a car and drive at 50 km/h alongside the train, the perceived speed of the train drops to 50 km/h as well. Let us consider a simple experiment whereby half the time you measure the train’s velocity while in the car and half the time you measure it outside the car, which we can call B, and we will define A as simply whether the train’s velocity is 100 km/h.
If we conduct this experiment, we would find that P(B)=50% as we defined it as such, and P(A)=50% as every time we randomly get in the car or get out the car, we also equally see the train’s velocity speed up or slow down. Additionally, P(A∩B)=0%, as there is never a point where we are in the car, and we see its velocity at 100 km/h. Even more so, P(A|B)=0% for the same reason. Clearly, then, P(A∩B)≠P(A)P(B). However, P(A|B)=P(A∩B)/P(B) as P(A∩B)/P(B)=0%/50%.
If we apply this same reasoning to Galilean relativity, we also find a statistical dependence between whether we are in the car or out of the car, and what we measure the train’s velocity to be. Does that then prove that the very act of getting into the car perturbs the train and changes its velocity? Of course not. Velocity is just a property of a system that depends upon point-of-view.
Recall that there is no more subtle measurement in quantum mechanics, so we tried to derive an observer effect indirectly by applying two measurements. The first measurement is a midway measurement between the two beam splitters, and the second measurement is at the two detectors. The only thing we varied is the midway measurement, and thus, in all cases, we are still observing the system. In none of these cases are we not observing the system, and thus we cannot derive any conclusions about whether our measurement is perturbing the system.
What we find is that if we look at the system differently, we see different results. This, on its own, does not demonstrate an observer effect. It only demonstrates that the outcome depends upon how we look at it. In other words, it is point-of-view dependent, or, in other words, context dependent. This is context in the philosophical sense, not in the hidden variable sense, which we can call contextual realism.
Contextual realism does indeed uphold, as opposed to hidden variable contextuality, the notion that the act of measurement does indeed reveal the particle as it is independent of observation. However, it is dependent upon the context of that observation, i.e., “how you look at it,” your point-of-view, a reference frame, a coordinate system, so on and so forth. Reality does exist independent of the observer, but not independent of context. You have to specify a coordinate system to assign properties to systems at all.
Again, we have already known this since Galilean relativity, as velocity is inherently relative to different points-of-view. Einstein extended this to both space and time itself. Quantum mechanics merely extends this to all variable states of particles. Whether a particle has a particular property or not, and what predictions you will make as to what its properties can be, depend upon your context, your point-of-view. When you make an observation, you are revealing the property of the system as it exists independent of the observer, but not independent of the context of your observation.
The point of view on the measurement process in quantum mechanics as a process of interaction between the instrument and the physical system is a kind of projectivism. This view places normative (the rules according to which a measurement is made, and therefore a measuring instrument as a measuring instrument) and real (quantum system) in the same category. This leads to an endless regression: more and more new observers have to be introduced to observe the previous observers. Regress stops if the above categorical difference is taken into account. In fact, the subject who uses the instrument does not impose his point of view from the outside (an outside view would mean that we classify the measuring instrument and the system above which the measurement is made into the same category), but is already in the context of the measurement, the point of view of the measuring instrument. Thus, the reduction of the wave function is not a physical process.
— Francois Igor Pris, “Contextual Realism and Quantum Mechanics”
Contextual Realism
The Meaning of Context
Clearly, there is not a one-to-one relation between the analogy of Galilean relativity and quantum mechanics. Both are, in a sense, relative, yet this requires some clarification on both the similarities and differences to avoid confusion.
First, what exactly is relative (i.e., context-dependent) in quantum mechanics? Two things: (1) the wave function, and (2) the definiteness of the states of particles. This is illustrated simply with the famous Wigner’s friend paradox. Wigner observes his friend making measurements as to whether a radioactive atom decays, which is probabilistic, and he can see his friend but cannot see her measurement results nor can he ask her for her results.
Wigner’s friend would describe the particle in a definite state that is exactly as she measured it, and thus also describe the system with a collapsed wave function. Wigner himself knows his friend is making measurements on the particle, so he would describe his friend as entangled (statistically correlated) with the particle, but could not assign a definite state to either of them, and thus would have to use an entangled wave function rather than a collapsed wave function.
What is not relative is the actual values of particles. If Wigner asks his friend what she saw, and he goes and measures it himself, he would expect those to match up. Two observers cannot look at the same electron, for example, and one sees it as spin up and the other as spin down. They will agree as to what the definite states of particles are. However, if they are isolated from one another and cannot communicate, when they describe the same system, they may not agree on which set of properties is definite and which set is indefinite.
Second, how exactly is the context changed? In the example of Galilean relativity, the reference frame is changed by getting in and out of a car, i.e., it is changed through acceleration. Clearly, accelerating is not going to have an effect on whether an electron has a definite spin value or not. How the context under which a system is observed is changed in quantum mechanics works a bit differently.
The context should be understood in terms of relations. A relation is just a kind of correlation. Imagine throwing a ball off a boat. Due to Newton’s third law, pushing the ball off the boat should result in it also pushing back, and thus the boat will move slightly in the opposite direction. If a person never saw the ball being thrown but saw the slight motion of the boat, they could infer the direction the ball was thrown. Why? Because they are correlated. The motion of the ball and the boat are related, and thus there is a relation between them.
Just like Galilean relativity, you first have to choose a coordinate system based on some other object. You can speak of the velocity of the train relative to the ground or relative to the car, but it is meaningless to speak of the velocity unto itself. It is not a property of the train itself but a relationship between the train and other objects, and thus to assign a velocity at all, you have to first select another object to be the basis of a coordinate system from which to describe the train.
In quantum mechanics, you also need to choose some object to be the basis of a coordinate system to describe other objects. We can call this the referent object. You then need to take into account the referent object’s context. The context should be understood merely to refer to the totality of its relations with its environment, or, at the bare minimum, the totality of its relations with the system in which you want to describe. This is what constitutes the context and thus the point-of-view in quantum mechanics, which properties of systems are relative to.
Answering the question of what leads to a change in context requires a discussion of the wave function.
The Meaning of the Wave Function
If the behavior and properties of systems are relative to a particular context, a particular point-of-view, then it logically follows that you have to take into account your context in order to predict what you will observe. This is merely what the wave function entails. It is an accounting of the observer’s perspective on a particular system based on what relations with the system they are trying to predict and describe.
When you make a measurement, you are not just revealing the particle’s state, but in order for you to “know” it, the particle is imprinting its state onto you, and thus you are forming a relation with it. This is typically not directly but indirect: the particle forms a relation with the detector (what’s the point of a detector if not to “reflect” and thus become correlated with the particle’s state) and then you look at the detector, which imprints that state into your brain.
Thus, your brain becomes correlated with the state of the particle, i.e. it forms a relation with it. This relation means that you can infer the state of the particle from your brain’s state. That’s how you know it in the first place! You could also infer your brain’s state from the state of the particle. If someone else observed the particle, and they knew you also looked at it, they would know that you know that state as well, and would thus come to know what you know.
Why am I talking about observers and brains if this is an observer-independent theory? The reasoning is rather simple, let’s go back to the example with Galilean relativity. It is technically meaningful to talk about the velocity of an object, like a train, in relation to an inanimate object like a rock. However, we typically don’t do this because we want to know what we will observe from our point-of-view. We don’t care about a rock’s point-of-view. Quite often, the observer chooses themselves as the basis of their coordinate system, as the referent object, and they do this implicitly without stating it as they are interested in what they will observe.
We choose ourselves as the referent object — as the basis of our coordinate system — implicitly, not because we have to, but because that’s what is most convenient. We want to know what we will observe from our point-of-view, so we end up taking into account our own context and thus our own relations with our environment.
The wave function is a description of our context in relation to the system we want to describe. Why is it a “wave”? Well, simply because some relations interfere with other relations. This is due to the principle of complementarity. If we measure a particle’s position, it interferes and, in a sense, negates, our relation with its momentum. Waves, of course, exhibit interference properties, and so this helps us describe how our relations with a system change over time.
The wave function thus does not describe a system’s properties as if the system is literally a wave. It is merely an accounting of the totality of our relations with that particular system, and thus it is merely a way of accounting for our context. It is, in a sense, an accounting of our “perspective” or “point-of-view” on that particular system. Once we take it into account, we can then make predictions as to what we may observe in the future.
If we make an observation, again, the act of observation necessarily, as already discussed, forms a relation between the particle’s state and ourselves. If context is based on relations, then, by definition, the act of observation changes our context, it changes our “point-of-view,” and therefore it logically follows you have to update your accounting of your context accordingly. In other words, you have to update the wave function. This is not due to disturbing some physical wave-like entity and causing it to collapse, but it is more like taring a scale, it is re-centering your coordinate system upon yourself after your point-of-view has changed.
This is how context is changed in quantum mechanics. It changes whenever you make an observation, as your relations with your environment — and thus your context — changes.
Unlike Galilean relativity where you can freely choose to get in the car and get out your car at your own leisure, in quantum mechanics, due to the probabilistic and nondeterministic nature of the theory, you cannot choose your point-of-view freely. It changes spontaneously every time you interact with something, and it is also not reversible. You cannot undo an observation. Part of your context you can indeed freely choose, such as the measurement setting, but as a result of measurement, your context is changed in a way you cannot control with a predetermined statistical correlation predicted by the wave function (and thus predetermined in that context), but the specific values of those measured particles are not predetermined.
According to our interpretation, at the moment of measurement there is not a splitting of the world or consciousness, but a transition to this or that context in which a certain quantum correlation is already predetermined. Outside the context, a certain correlation is not predetermined, only the correlation itself is predetermined. Moving into one context or another corresponds to the choice of coordinate system (point of view); it is not a physical process. In that sense, the word “transition” isn’t exactly good. An observer simply discovers that he or she is in a certain context, within a certain point of view (in this case, unlike in classical physics, he or she cannot choose his or her context and cannot return to the original position). If the “coordinate system” is fixed, the correlated value of the physical quantity is fixed. So the quantum correlation is “coordinate”. It is coordinate both in the sense of the initial choice of the “coordinate system” and in the sense of the coordinate dependence of correlated physical quantities at a fixed choice of the initial coordinate system.
— Francois-Igor Pris, “Contextual Realism and Quantum Mechanics”
Hence, it is clear, then, why Wigner and his friend describe the same system differently: they occupy different contexts as they have different relations to the same system, and thus have a different point-of-view. When they account for their unique contexts using the wave function, they arrive at different wave functions. Wigner’s friend will also view the particle as having a definite value and Wigner will view it as having an indefinite value, at least until he talks to his friend or measures it himself.
When Wigner does measure it himself, he is indeed revealing its value without perturbing it, he is seeing it exactly as it is independent of whether a conscious observer was there to look at it. However, he is not seeing it independent of the context of his observation. To assign a system properties as well as to predict its behavior, you have to specify a coordinate system, a context, as these things both depend upon that context. Yet, there is no special role here for conscious observers. It is just relative to context.
The correlated quantum events are not autonomous, but they are determined in the context of their observation. Independently from the means of their identification, there are no events. The reduction of a wave function in the «process of measurement» is not a real physical process, requiring an explanation, but a move to a context of measurement of a concrete value of a physical quantity. Respectively, the measurement is not a physical interaction leading to a change in the state of a system, but the identification of a contextual physical reality. That is, in a sense, in measuring (always in a context), one identifies just the fragment of reality where the (quantum) correlation takes place. As the elements of reality, the correlated events do not arise; they are. Only their identifications do arise.
— Francois-Igor Pris, “The Real Meaning of Quantum Mechanics”
Locality
This interpretation also allows us to interpret quantum entanglement entirely locally. Consider the famous EPR paradox. If Alice and Bob are sent entangled quantum coins that are guaranteed such that they will be opposite when observed, then if Alice looks at her coin and sees it is heads, she knows Bob’s will be tails, and vice versa.
This is often stated to be “spooky action at a distance” because Alice, by measuring her coin, collapses her wave function, allowing her to predict Bob’s outcome. If the wave function is a description of the physical state of a system, then yes, this would be “spooky” indeed, as it would imply, there is a giant invisible wave stretching between Alice and Bob that when Alice tries to measure her particle, she perturbs it, causing it to collapse into two particles for both Alice and Bob in that very moment, even if they are thousands of miles separated from one another.
Yet, again, the wave function is not a description of the state of a system. It is an accounting of the context in which one occupies, their point-of-view, in order to predict the state of particles. This prediction implies an observation, and thus an interaction, from their own context, i.e., from their own point-of-view. They are not describing anything happening to Bob’s particle, but updating their prediction as to what Bob’s particle will be if they were to go measure it themselves in the future.
Indeed, even in Bell tests, you can only demonstrate violations of Bell inequalities by first having the two particles come together locally in order to become correlated with one another, sending the two particles off to distant observers, and then bringing the particles back together locally to compare results. The apparent “spooky action at a distance” only shows up if you make certain metaphysical assumptions after this entirely local experiment that are not justified, such as the wave function being a description of the state of a system.
The paper “The notion of locality in relational quantum mechanics” singles out five notions of locality that are not equivalent at all, although closely related: 1) signals cannot spread faster than the speed of light; 2) the spatial distance of causes and effects from each other is limited by the speed of light; 3) quantum systems, separated by a space-like interval, do not affect each other; 4) interactions are possible only at the same point of spacetime; 5) local variables relating to areas of space separated by space-like interval commute. Our interpretation of quantum mechanics within the framework of contextual realism is compatible with all these notions of locality.
— Francois-Igor Pris, “Contextual Realism and Quantum Mechanics”
It is tempting to think of the wave function as describing the ontological state of a system, but such a view just cannot reproduce the findings of quantum mechanics. There are various no-go theorems which demonstrate this. See this paper, this paper, and this paper. There is a myth that you are forced to accept the wave function represents the ontological state of a system due to the PBR theorem, however, the PBR theorem is only applicable to classical epistemic interpretations. Such a contextual realist view is relational, and in no way classical, and thus the PBR theorem does not apply.
