🌊 The Importance of Context, or Why Watching from the Shore Cannot Replace Sailing the Dirac Sea

The Kochen–Specker Theorem and the Limits of a Pre-Determined Quantum Reality

In our previous voyage across the Dirac Sea, we explored Bell’s inequality. We saw that nature does not allow the simple picture in which distant quantum systems merely carry a pre-arranged set of local values that measurement subsequently reveals. The experimental violation of Bell inequalities placed severe constraints on this classical picture of local realism.

But the problem does not end there.

Bell took us toward distant shores: two systems, two measurements, spatial separation, and correlations that local models cannot reproduce. The Kochen–Specker theorem asks an even more uncomfortable question. Do we actually need two distant systems for the problem to arise?

The answer is no.

The difficulty already appears within a single quantum system as soon as we attempt to assign values in advance to all its possible observables, independently of the way in which they will be measured together. This is where the concept that will be central to this voyage appears: context.

The Kochen–Specker theorem is one of the central results in the foundations of quantum mechanics. In modern language, it demonstrates a conflict between quantum theory and models in which the outcome of a measurement does not depend on which other compatible measurements are performed alongside it. This conflict is now known as quantum contextuality.

When an Observable Becomes Part of a Context

We will not return here to the question of what an operator is or why a physical observable in quantum mechanics is represented by a Hermitian operator. For this discussion, at least a basic familiarity with the operator formalism must be assumed.

Let a quantum state be denoted by

and let two observables be represented by the operators

If the operators commute,

then there are conditions under which their values can be considered and measured jointly. In the language of contextuality, such a set of mutually compatible observables forms a measurement context.

In classical physics, there is usually nothing disturbing about this. A body has mass, position, velocity, angular momentum and other physical quantities regardless of the order in which we decide to investigate them. Measurement may be imperfect and may disturb the system, but we naturally assume that properties do not arise merely because we changed the rest of the experimental apparatus.

In quantum mechanics, this assumption is no longer innocent.

The Kochen–Specker question is not simply: which value will we obtain? The deeper question is: can we assign values in advance to all possible observables in such a way that each observable retains the same value regardless of the context in which it is measured?

At first sight, this seems like a perfectly reasonable requirement. If a physical quantity truly represents a property of a system, why should its value depend on which other compatible quantities we have chosen to measure alongside it?

This is precisely where the problem begins.

Spin-1: A Small Construction with Large Consequences

The clearest way to see the essence of the Kochen–Specker theorem is through a spin-1 system.

For spin s = 1, we have the familiar relation

The same relation can be written for any three mutually orthogonal directions x, y and z.

For a spin-1 particle, measurement of a spin component gives

so the square of the corresponding component has only two possible values:

or

If, for convenience, we set ℏ² = 1, we obtain

while at the same time

must hold.

Therefore, for every orthogonal triple of directions, the possible values must be some permutation of

Two ones and one zero.

At this point there is no paradox. In fact, everything still looks almost classical. We may try the following: assign in advance either 0 or 1 to every possible direction in space. When we later choose three mutually orthogonal directions, we simply read the values that were already there. The only requirement is that every such orthogonal triple contain two ones and one zero.

In other words, we are trying to construct a large pre-prepared table of answers. The quantum system would not have to “decide” anything during measurement. Everything would already be written down.

The Impossible Coloring of Space

The problem can now be transformed into a geometric task.

Imagine directions in three-dimensional space as points to which we assign two colors. One color represents the value 0, the other the value 1. The rule is simple: for every orthogonal triple, exactly one direction must carry the value 0 and the other two the value 1.

For a single triple this is easy. For several triples it is still easy. But different orthogonal triples share the same directions. A direction that has already been assigned a value in one context must retain that value when it appears in another. If we demand noncontextuality, we cannot assign it 0 in one case and 1 in another merely because the other two directions in the experiment have changed.

This is where the network begins to close upon itself.

Kochen and Specker showed that there exists a finite set of directions for which this task cannot be solved. Their original 1967 construction was large, and later much more economical configurations were found, but the logic remains the same: there is no global assignment of 0 and 1 that simultaneously satisfies all the orthogonality requirements.

This is not a problem of our ignorance. It is not a problem of an insufficiently clever algorithm. It is not that we simply have not yet discovered the “true” values. Such a global noncontextual assignment is mathematically impossible.

This is precisely why the Kochen–Specker theorem is far more uncomfortable than it first appears.

What Has Actually Been Lost?

At this point we should stop for a moment, because it is very easy to move from physics into metaphysics.

A popular simplification would be:

“The Kochen–Specker theorem proves that properties do not exist until we measure them.”

No.

The theorem does not prove this. Nor does it prove that physical reality does not exist independently of human beings, that consciousness creates the outcome of measurement, or that no hidden-variable theory can exist.

Its claim is narrower but much more precise: we cannot assign context-independent values in advance to all quantum observables while simultaneously preserving the functional relations required by quantum theory.

That is quite enough.

The classical picture in which a system carries an enormous table containing every possible answer, while the experimenter merely chooses which entry to read, is no longer sustainable. But this does not imply that no deeper reality exists. It implies that, if such a reality exists, it cannot be organized in such a simple way.

Bell and Kochen–Specker: Two Cracks in the Same Shore

The historical connection between Bell and the problem of contextuality is particularly interesting. In 1966, one year before the famous paper by Kochen and Specker, Bell revisited the hidden-variable problem and showed how important it is to examine carefully the assumptions hidden inside earlier no-go arguments.

For this reason, Bell and Kochen–Specker frequently appear close together in the history of the foundations of quantum mechanics. Nevertheless, their messages should not be identified with one another.

With Bell’s inequality we have distant systems, different local measurement choices and correlations that exceed the limits of local hidden-variable models. With Kochen–Specker, large spatial separation is unnecessary. The problem can arise within a single system. The central issue is no longer primarily locality, but the possibility of assigning the same pre-existing value to an observable in every compatible measurement context.

Bell nonlocality can therefore be regarded as a special manifestation of the broader problem of contextuality, but the Bell and Kochen–Specker theorems are not the same result.

Bell revealed a crack between distant shores. Kochen–Specker showed that the crack also exists on the land itself.

What Does Context Mean Physically?

This is where mathematics begins to raise philosophical questions.

If the result of an observable cannot be imagined as a single pre-existing value entirely independent of the set of other compatible observables with which it is measured, then a natural question arises: what exactly do we mean by a physical property?

Classical intuition tells us that measurement should reveal something that already existed. A thermometer reads temperature, a voltmeter voltage, an ammeter current. Of course, every real instrument affects the system to some degree, but conceptually we separate the property itself from the procedure by which it is measured.

The Kochen–Specker theorem shows that this concept cannot be transferred without restriction into the quantum world. This still does not mean that the instrument “invents” the result. The distinction is subtler.

Perhaps the very idea that every possible observable must possess one globally defined value before we specify the physical context of measurement is a demand that nature is simply not obliged to satisfy. That is a very different claim from saying that reality does not exist.

Bohm Once Again Warns Us Not to Conclude Too Much

Bohmian mechanics provides an especially useful control example here.

When the Kochen–Specker theorem is discussed, it is sometimes presented as though it had finally eliminated every possible form of hidden variables.

It has not.

Bohmian theory contains additional variables. However, the outcome of a particular measurement can depend on the entire experimental arrangement. Bohmian mechanics is therefore contextual. This allows it to evade the restriction imposed by the Kochen–Specker theorem.

Of course, Bohmian mechanics pays other prices, among them the explicit nonlocality that we already encountered in our previous voyage through Bell’s inequality. For our present discussion, however, it is important because it blocks one misleading interpretation: Kochen–Specker does not destroy the very idea of a deeper level of reality. It destroys the noncontextual way in which we would most naturally like to imagine that deeper level by analogy with classical physics.

This is an important distinction.

When a Mathematical Theorem Enters the Laboratory

Here we arrive at another slippery shore.

The Kochen–Specker theorem is a mathematical theorem. In that sense, its validity does not await confirmation from some future experiment. Experiments can, however, test whether physical systems display the operational consequences of quantum contextuality.

That is not exactly the same question.

A real experiment no longer consists of ideal operators drawn in chalk on a blackboard. It contains detectors, finite precision, sequential measurements, imperfect compatibility, measurement back-action and statistical processing.

For this reason, the modern theory of contextuality has developed various experimentally testable inequalities and constructions, including the KCBS approach, Peres–Mermin configurations and other measurement schemes capable of distinguishing quantum predictions from particular classes of noncontextual models.

And this introduces further questions. How compatible are two real measurements in practice? How much disturbance does one measurement introduce into the next? Does an experimental procedure truly realize the ideal context assumed by the mathematical construction?

There are also modern techniques based on very weak interactions between a system and a measuring device in order to reduce the influence of the measurement itself. But this already opens another debate: what exactly does such a “weak” measurement measure, and how much ontological meaning are we entitled to assign to the resulting value?

That is a large enough subject to deserve another voyage of its own. For now, it is enough to leave a lighthouse on that shore.

From Wigner’s Friend to Relational Reality

For readers who have followed the Dirac Sea series for some time, this problem will probably sound familiar.

We have already encountered it through Wigner’s friend, relational quantum mechanics, QBism, quantum Darwinism and different interpretations of measurement.

The Kochen–Specker theorem does not choose a winner among these approaches. It does not say that relational quantum mechanics is correct. It does not say that QBism is the final answer. It does not confirm the many-worlds interpretation. It does not refute Bohm.

But it helps explain why such different approaches arise in the first place. Each of them, in a different way, attempts to answer the same uncomfortable question: what constitutes a physical fact if the world cannot be imagined as a pre-filled catalogue of all possible quantum values?

Relational quantum mechanics emphasizes relations between physical systems. QBism emphasizes the experience and expectations of an agent. Quantum Darwinism attempts to explain how certain information becomes stable, redundant and accessible to many observers, producing what we experience as objective classical reality. Bohm retains a strong ontology, but accepts contextuality and nonlocality. MWI eliminates a unique collapse, but at the price of the ontology of branching worlds that we have already discussed elsewhere.

The Kochen–Specker theorem does not decide between them. It merely removes one comfortable place where we might have wished to remain.

What the Kochen–Specker Theorem Does Not Prove

It is therefore worth drawing the boundary very clearly.

It does not prove that the world ceases to exist when nobody observes it. It does not prove that human consciousness creates reality. It does not prove that no physical quantity possesses a value before measurement. It does not prove that wave-function collapse is an objective physical process. It does not choose Copenhagen over Bohm, QBism over MWI, or relational quantum mechanics over quantum Darwinism. Nor does it turn physics into subjectivism.

What it removes is a much more specific possibility: we cannot imagine that all quantum observables simultaneously carry predetermined, context-independent values and then expect those values to reproduce the structure of quantum theory.

This may sound more modest than the claim that “the observer creates reality.” But precisely for that reason it is more serious. It does not depend on metaphor. It depends on mathematics.

Watching from the Shore Is Not the Same as Sailing

In the classical world, we can imagine an ideal cartographer who never leaves the harbor. He knows the terrain, the depth of the sea, the direction of the wind and the location of the reefs. Given enough information, he could in principle calculate the complete route of a ship in advance. The wave that the ship will encounter at a particular point is simply part of a physical world that already existed before the ship arrived there.

The Kochen–Specker theorem tells us that we cannot simply transfer this picture to the Dirac Sea. We cannot remain safely anchored in the harbor and write down in advance the result of every possible measurement along every possible route, as though all those values simultaneously existed in one great book of navigation.

The ship must choose a route. It must set sail. And along that particular path it must measure the wave that actually strikes it. This does not mean that the sea was unreal before we sailed. Nor does it mean that the ship created the wave by observing it.

But it does mean that we cannot assign to all possible routes one single context-independent collection of results in advance and expect it to remain consistent when we attempt to combine all those routes into one map. We may know the laws of the sea. We may know the Hilbert space. We may know the allowed eigenvalues and the relations between compatible observables. But the complete route does not exist as a simple list of pre-written answers that can be read from the shore.

In our previous voyage, Bell showed us that distant shores are not connected in the way local classical intuition would prefer. Kochen–Specker goes one step further. Now even the view from a single shore is not enough, because in the Dirac Sea the question we ask of nature cannot be separated from the context in which we ask it.

And perhaps this is the deepest message of the theorem.

It is not enough to ask what the world is like. We must know the context in which we asked the question. 🌊⚛️