Pseudotensor

In physics, tensors offer a universal language for describing nature's laws, ensuring they remain consistent regardless of our chosen perspective or coordinate system. However, not all physically significant quantities adhere to the strict transformation rules of a true tensor. This raises a crucial question: What are these 'almost tensors,' and what do they reveal about the universe? This article confronts this knowledge gap, demonstrating that these objects, known as pseudotensors and tensor densities, are not mathematical flaws but essential tools for describing a world rich with asymmetry and chirality. The first chapter, ​​Principles and Mechanisms​​, will demystify these objects by examining their unique transformation laws and contrasting them with true tensors. Following this, the chapter on ​​Applications and Interdisciplinary Connections​​ will showcase their indispensable role in phenomena ranging from the optical properties of crystals to the very energy of spacetime in Einstein's theory of gravity.

In our journey to describe the universe, physicists are like cartographers mapping a vast, unseen territory. Our maps are coordinate systems, and the laws of physics are the features—the mountains, rivers, and cities—that must look the same no matter which map we use. Tensors are the language we invented for this task. They are geometric objects whose essence remains unchanged by our choice of coordinates. A law written in the language of tensors is a true law of nature.

But what happens when we encounter quantities that almost fit this perfect description? What if they change in a predictable, yet not-quite-tensor way? Do we discard them as mathematical artifacts? Absolutely not! As we shall see, these "almost tensors" are not flaws in our description; they are clues, revealing deeper and more subtle symmetries of the universe itself.

Let’s first be clear about what we expect from a true tensor. When we switch from one coordinate system to another, a tensor's components transform in a clean, "linear" fashion. Each new component is a neat combination of the old components, weighted by factors that are just the rates of change (first derivatives) of the old coordinates with respect to the new ones.

Consider, for contrast, the ​​Christoffel symbols​​, $\Gamma^k_{ij}$, which you may have encountered in studies of curved space. They are essential for describing how vectors change from point to point. At first glance, their transformation law looks a bit like a tensor's, but it's contaminated by an extra piece—an additive term involving second derivatives of the coordinates. This extra term is like a smudge on our lens; it depends on the coordinate system itself, not just the object we are trying to look at. Because of this ​​inhomogeneous transformation law​​, the Christoffel symbols are definitively not tensors. Their value can be made zero at a single point just by choosing a clever coordinate system (a locally "flat" one), something you could never do with a real physical field like an electric field.

This failure is instructive. We are interested in objects whose transformation laws are homogeneous—meaning if the object is zero in one frame, it's zero in all frames. The objects we will now explore, tensor densities and pseudotensors, pass this test. Their deviation from the tensor ideal is not an unwanted additive piece, but a clean, multiplicative factor.

There are two main ways a quantity can be an "almost tensor" while still being physically meaningful. Both involve a special multiplier in their transformation law that depends on the Jacobian matrix of the coordinate change, let's call it $L^i_j = \frac{\partial x'^{i}}{\partial x^{j}}$. The determinant of this matrix, $\det(L)$, tells us how the transformation scales volumes and whether it flips orientation.

1. ​​Tensor Densities​​: These objects are sensitive to the scale or "volume" of our coordinates. Their transformation law includes a factor of $(\det(L))^W$, where the exponent $W$ is a number called the ​​weight​​. A true tensor is simply a tensor density of weight $W=0$.

2. ​​Pseudotensors​​: These objects are sensitive to the handedness or ​​orientation​​ of our coordinates. Their transformation law includes a factor of $\text{sgn}(\det(L))$, which is just the sign ($+1$ or $-1$) of the determinant.

Let's unpack these two fascinating characters one by one.

Imagine trying to measure the density of a fog. You might define it as the amount of water in a cubic meter box. What happens if you change your coordinate system, stretching it so your unit "box" is now twice as large in volume? The amount of water in the box doubles, but if you want the density (mass per unit volume), its numerical value must be halved. This quantity, density, is a perfect example of a scalar density of weight $W=-1$.

This is the essence of a ​​tensor density​​: its components change to compensate for the stretching or shrinking of the coordinate volume element, represented by $\det(L)$.

A truly fundamental example is the famous ​​Levi-Civita symbol​​, $\epsilon_{i_1 \dots i_n}$. In three dimensions, you know it as $\epsilon_{ijk}$, which is $+1$ for $(1,2,3)$ and its cyclic permutations, $-1$ for anti-cyclic ones like $(1,3,2)$, and $0$ if any indices repeat. Its definition seems fixed, but this rigidity is exactly why it cannot be a true tensor. When you change coordinates, it transforms as a covariant tensor density of weight $W=-1$. It's as if its very definition is tied to the coordinate labels "1, 2, 3" rather than the underlying space.

Happily, there's a beautiful simplicity to how these weights combine. If you multiply a tensor density of weight $W_1$ with another of weight $W_2$, the resulting object is a tensor density of weight $W_1 + W_2$. This leads to a lovely trick: if you have a tensor density of weight $W_A$, you can make it part of a true tensor (weight 0) simply by contracting it with another density of weight $W_B = -W_A$. The weights cancel out, and you're left with a geometrically pure object!

This machinery works beautifully. We can perform complex calculations, like finding the components of a tensor density in a weird, curved coordinate system, and the weight $W$ will show up exactly where it should, scaling the result by the appropriate power of the Jacobian determinant. We can even define calculus for these objects; the covariant derivative of a tensor density can be defined, and its divergence turns out to be a vector density of the same weight, ready to be used in physical laws.

Now for the other character in our story: the pseudotensor. Its transformation rule picks up a factor of $\text{sgn}(\det(L))$. What does this mean? $\det(L)$ is positive for rotations (which preserve "handedness") and negative for reflections (which swap left and right hands). So, a pseudotensor transforms exactly like a regular tensor under rotations, but it flips its sign under a reflection. It is an object that knows the difference between a system and its mirror image.

The most famous example is the humble ​​cross product​​ in 3D. The vector $\vec{C} = \vec{A} \times \vec{B}$ is defined by the right-hand rule. Now, imagine this operation in a mirror. The reflections of $\vec{A}$ and $\vec{B}$ obey the laws of optics. But if you apply the right-hand rule to these reflected vectors in the mirror world, the resulting vector points opposite to the reflection of the original vector $\vec{C}$. The cross product vector is a ​​pseudovector​​ (or axial vector), a pseudotensor of rank 1.

The agent responsible for this behavior is, once again, the Levi-Civita symbol. When we construct an object like $A_{jk} = \epsilon_{ijk} v^i$ from a true vector $v^i$, the result is not a true tensor, but a rank-2 pseudotensor.

This is where a crucial distinction comes into play from problem. The Levi-Civita symbol $\epsilon_{ijk}$ is a tensor density. But if we combine it with the metric tensor $g_{ij}$ to form the ​​Levi-Civita tensor​​, $\tilde{\epsilon}_{i_1\dots i_n} = \sqrt{|\det(g)|} \epsilon_{i_1\dots i_n}$, something magical happens. The $\sqrt{|\det(g)|}$ factor is precisely what's needed to cancel the "density" behavior of the symbol, leaving behind only the "pseudo" behavior. This object, the Levi-Civita tensor, transforms as a true pseudotensor. It is the mathematical tool for building orientation-dependent quantities in physics.

A beautiful application of this is the ​​Hodge dual​​, an operation often used in electromagnetism. This operation intrinsically uses the Levi-Civita pseudotensor to turn an antisymmetric tensor into another one. Consequently, the Hodge dual of a true tensor is always a pseudotensor. This explains why a physicist performing calculations might find that "dualizing then transforming" gives a different answer from "transforming then dualizing" if the transformation involves a reflection—the missing sign is the tell-tale signature of a pseudotensor!

The "algebra of parity" is as simple as the algebra of weights. A pseudotensor times a pseudotensor gives a true tensor, because $(\text{sgn}(\det(L)))^2 = 1$, washing away the orientation dependence. A pseudotensor times a true tensor remains a pseudotensor. Contracting the indices of a mixed pseudotensor can produce a ​​pseudoscalar​​—a single number that is invariant under rotations but flips its sign under reflection.

What about other properties, like symmetry? If you have a symmetric pseudotensor, $P^{ij} = P^{ji}$, and you view it in a mirror, all its components flip sign ($P'^{ij} = -P^{ij}$). However, the symmetry property itself is preserved: it is still true that $P'^{ij} = P'^{ji}$.

Why do we care about these subtleties? Why not just stick to our perfect, true tensors? Because Nature herself is not always perfectly symmetric.

Physical quantities like ​​angular momentum​​, ​​torque​​, and the ​​magnetic field​​ are all pseudovectors. You cannot describe rotation correctly without them. The fact that they behave differently in a mirror is not a mathematical curiosity; it's a deep fact about their rotational nature.

Even more profoundly, some fundamental laws of physics are not mirror-symmetric. The ​​weak nuclear force​​, responsible for certain types of radioactive decay, famously violates parity conservation. The universe, at a fundamental level, has a "handedness." To describe such phenomena, pseudotensors are not just helpful; they are essential.

In Einstein's theory of General Relativity, the energy and momentum of the gravitational field itself is described not by a true tensor, but by a pseudotensor. This is connected to the profound idea that gravitational energy cannot be confined to a specific point in space.

So, the existence of tensor densities and pseudotensors is not a complication to be swept under the rug. It is a vital part of the physicist's toolkit, a language that allows us to describe the full, rich, and sometimes wonderfully asymmetric reality of the universe. They are a testament to the fact that in physics, even a failure to conform to a simple ideal can be a doorway to a deeper understanding.

Alright, we have spent some time getting acquainted with these curious objects called pseudotensors. We’ve seen how they behave in a mirror—transforming almost like their well-behaved cousins, the true tensors, but with a crucial twist, a flip of sign under reflection. You might be tempted to think this is just a mathematical subtlety, a bit of bookkeeping for pedantic physicists. Nothing could be further from the truth. In fact, the distinction between a tensor and a pseudotensor is not a complication to be endured, but a profound clue from nature. It is the language physics uses to describe a world that is not perfectly symmetric, a world filled with handedness, chirality, and intricate couplings between seemingly unrelated phenomena.

Let’s now go on a journey, from the heart of a crystal to the fabric of spacetime itself, to see where these pseudotensors live and what secrets they reveal about the universe.

A perfect, empty space is the same in all directions and has a perfect mirror symmetry. But the moment you put matter in it, you start breaking that symmetry. A crystal is a wonderful example. It has a beautiful, repeating lattice, but that lattice has its own specific symmetries—it might look the same if you rotate it by 90 degrees, but not by 75 degrees. More importantly, it might not look the same as its reflection. This lack of mirror symmetry, or inversion symmetry, is where pseudotensors take center stage.

The Twist of Light

Have you ever wondered why some crystals, like quartz, can rotate the polarization of light that passes through them? This phenomenon, called optical activity, is a direct signature of the crystal's "handedness," or chirality. The effect is described by relating the direction the light is traveling, say along a direction given by the polar vector $\mathbf{n}$, to a quantity called the gyration vector $\mathbf{G}$, which is an axial vector. The relationship is linear: $G_i = g_{ij} n_j$.

Now, think about this in a mirror. The direction vector $\mathbf{n}$ flips, like any arrow. But the gyration vector $\mathbf{G}$, being axial (think of it like the axis of a spinning top), does not. For the law of physics to hold in the mirror world, the object connecting them, the gyration tensor $g_{ij}$, must have a screw loose—it must be a pseudotensor. Its job is to absorb that minus sign from the reflection. Furthermore, the specific atomic arrangement of the crystal dictates the form of this tensor. For example, in a crystal with a four-fold rotational symmetry (like the point group $D_4$), a careful analysis of the symmetry requirements shows that this $3 \times 3$ tensor, which could have nine independent components, is constrained to have only two. Symmetry is a powerful taskmaster!

The Marriage of Magnetism and Electricity

The connections can be even more surprising. Imagine taking a special material, applying a magnetic field to it, and watching an electric polarization appear out of nowhere. This is the linear magnetoelectric effect, described by the simple-looking equation $P_i = \alpha_{ij} B_j$. Here $\mathbf{P}$ is the electric polarization, a true polar vector that flips its direction in a mirror. But the magnetic field $\mathbf{B}$ is an axial vector; it represents a circulation, a spin, and its direction doesn't flip in a mirror.

What, then, is the character of the coupling tensor $\alpha_{ij}$? For the equation to remain true in a mirror, where the left side flips sign but the right side doesn't (since $\mathbf{B}$ is unchanged), the tensor $\alpha_{ij}$ must itself flip its sign. It must be a pseudotensor. But there's a deeper consequence. A material property, like $\alpha_{ij}$, must respect the symmetries of the material itself. If the material's crystal structure is "centrosymmetric"—meaning it is identical to its own mirror image—then its properties cannot change under reflection. But we just saw that $\alpha_{ij}$ must change sign. The only way to satisfy both conditions is for $\alpha_{ij}$ to be zero! Therefore, the magnetoelectric effect is strictly forbidden in any crystal that possesses inversion symmetry. It can only exist in materials that are fundamentally chiral at the atomic level, where the very structure of the material provides the necessary broken symmetry for this "forbidden" marriage of magnetism and electricity to occur.

This is a general rule: pseudotensors are the agents of phenomena that couple polar and axial quantities, and these phenomena are hallmarks of systems that lack mirror symmetry. We see this pattern again and again, in complex effects like the elasto-gyration effect (where mechanical stress, a polar tensor, induces optical gyration, an axial tensor effect) or in various forms of non-linear optics induced by magnetic fields. In each case, a tell-tale pseudotensor appears in the constitutive equations, acting as a fingerprint of broken inversion symmetry. The specific form and number of non-zero components of these tensors are always dictated by the crystal's point group, a task for which the mathematical tools of group theory are exquisitely suited.

The Secret Dance of Spins

Perhaps one of the most elegant and important roles for a pseudotensor in modern physics is in describing the Dzyaloshinskii-Moriya (DM) interaction. In some magnetic materials, again those without inversion symmetry, neighboring magnetic spins don't just want to align parallel or anti-parallel. They prefer to be slightly canted with respect to each other, leading to beautiful, swirling spiral or vortex-like patterns called skyrmions. This tendency comes from an energy term in the system's Hamiltonian of the form $f_{\mathrm{DM}} = D (\mathbf{m} \cdot (\nabla \times \mathbf{m}))$, where $\mathbf{m}$ is the magnetization direction.

Let's look at this term. The magnetization $\mathbf{m}$ is an axial vector. The curl operator, $\nabla \times$, turns it into another axial vector. The dot product of these two axial vectors results in a quantity that is a pseudoscalar—a single number that is invariant under rotations, but flips its sign under inversion. For the total energy of the crystal to be a true scalar (as it must be), this pseudoscalar energy density can only exist if the crystal structure itself is chiral, providing a context where a pseudoscalar is an allowed invariant. The tensor lurking behind this interaction, connecting one spin to the spatial change in another, is none other than the fundamental isotropic pseudotensor, the Levi-Civita symbol $\epsilon_{ijk}$. The DM interaction is the key to creating and manipulating magnetic skyrmions, which are at the heart of proposals for next-generation data storage and spintronic devices.

The utility of pseudotensors is not confined to the exotic world of crystals. They appear in much more general contexts, enforcing the rules of consistency for the fundamental laws of physics.

Pierre Curie first articulated a wonderfully simple but powerful idea: in an isotropic system, an effect cannot have a lower symmetry than its cause. You can't get a directed flow (a vector) from a perfectly uniform pressure change (a scalar). This principle, when applied to the thermodynamics of irreversible processes, forbids many potential couplings between physical processes. For instance, in an ordinary fluid, a chemical reaction (a scalar process) cannot directly cause a heat flow (a vector process).

But these rules can be bent if you introduce a new element that breaks the symmetry. An external magnetic field $\mathbf{B}$ is a perfect example. Since it is an axial vector, it can mediate couplings that were previously forbidden. For instance, it allows the Hall effect, where an electric field (polar vector force) drives an electric current (polar vector flux) that has a component perpendicular to the field itself. The coupling looks like $\mathbf{J} \propto \mathbf{B} \times \mathbf{E}$. This cross product is only possible because the magnetic field, an axial vector, is present to break the mirror symmetry of the setup.

Sometimes, the requirement that a physical law must be consistent under all circumstances forces a quantity to be a pseudotensor. In advanced theories of continuum mechanics that account for the rotational freedom of microscopic particles within a material (so-called micropolar theories), the equation for the balance of moments involves a term with the Levi-Civita symbol. For the entire equation to transform correctly under a reflection, the "couple-stress" tensor, which describes the torques within the material, is forced to be a pseudotensor. The laws of physics themselves pick out the correct mathematical object for the job.

We end our journey with the most mind-bending application of all—one that touches upon the very nature of energy and gravity. We all learn that energy is conserved. In classical and special relativistic physics, this is a local law. The flow of energy is described by a stress-energy tensor $T^{\mu\nu}$, and its conservation is expressed by the elegant equation $\partial_\nu T^{\mu\nu} = 0$.

When Einstein developed General Relativity, he ran into a famous and profound problem. The equation for matter in a gravitational field becomes $\nabla_\nu T^{\mu\nu} = 0$, where $\nabla_\nu$ is the covariant derivative that knows about the curvature of spacetime. Because of the extra terms in this derivative, this equation no longer corresponds to a simple conservation law. It describes the exchange of energy and momentum between matter and the gravitational field. So, you ask, can we just add a term for the energy of the gravitational field itself, say $t^{\mu\nu}$, and get a total conservation law?

The answer is yes and no. You can define such objects, called gravitational stress-energy pseudotensors. But why a pseudotensor? The reason lies in the heart of General Relativity: the Equivalence Principle. This principle states that at any point in spacetime, you can choose a reference frame (a freely falling elevator) where the effects of gravity locally vanish. If the energy of the gravitational field were described by a true tensor, you couldn't just make it disappear by changing your coordinate system. A tensor that is zero in one frame is zero in all frames. Since we can make the local gravitational field vanish, its energy cannot be represented by a tensor. Any object we construct to represent it, like the Landau-Lifshitz pseudotensor, must be non-tensorial. It will have a non-zero value in one coordinate system and a zero value in another at the same point. This is the very definition of a coordinate-dependent artifact, not a physically "real" local entity.

This is a stunning conclusion. There is no well-defined, local energy density for the gravitational field. The energy of gravity is non-local; it's stored in the global curvature of spacetime. The pseudotensor is our bookkeeping tool to handle this subtle situation, allowing us to recover a meaningful, globally conserved energy for an isolated system, but at the cost of giving up the idea of pinpointing the location of gravitational energy.

From the twist of light in a chiral crystal to the elusive energy of the cosmos, pseudotensors are far more than a mathematical curiosity. They are nature's way of telling us about asymmetry, about chirality, and about the deep and sometimes subtle rules that govern the interplay between matter, fields, and the very geometry of spacetime.