A pseudovector, or axial vector, is a quantity in three-dimensional Euclidean space that transforms like an ordinary vector under rotations, but differently under reflections. If an orthogonal transformation is represented by a matrix , an ordinary, or polar, vector transforms as , whereas a pseudovector transforms as . The extra determinant factor distinguishes quantities associated with oriented planes or rotation from quantities associated with displacement. Common examples include angular momentum, torque, angular velocity, and the magnetic field.
Transformation law
Let be a real orthogonal matrix, so that
Its determinant is either or . The defining transformation law for a pseudovector is
There are two cases:
- Proper rotations, with , preserve orientation. Pseudovectors and polar vectors transform identically.
- Improper orthogonal transformations, with , reverse orientation. A pseudovector acquires an additional minus sign relative to a polar vector.
Thus, observing only rotations cannot distinguish the two types of quantity.
A particularly important transformation is spatial inversion, also called parity:
In three dimensions, , so
A polar vector reverses under spatial inversion, whereas a pseudovector remains unchanged.
This does not mean that a pseudovector remains unchanged under every reflection. For reflection across the -plane,
and consequently
The component perpendicular to the mirror plane is unchanged, while the components parallel to it reverse.
These formulas can describe an active transformation of a physical configuration or a passive change between orthonormal coordinate systems, provided the convention for is used consistently.
Cross products and handedness
The standard source of pseudovectors is the cross product of two polar vectors. For any orthogonal matrix ,
Therefore, if and are polar vectors, their cross product is axial.
For example, take
Reflecting across the -plane gives
so
An ordinary vector initially directed along the positive -axis would instead remain unchanged under that reflection.
The distinction arises because the usual cross product depends on a choice of handedness. Its direction is conventionally assigned by the right-hand rule. Reversing the chosen orientation reverses the cross-product operation, even though the underlying plane and its sense of rotation can be described without selecting a normal arrow.
The word “pseudo” does not mean that the quantity is fictitious or mathematically invalid. It identifies a different transformation law.
Physical examples
Angular momentum and torque
In classical mechanics, the orbital angular momentum of a particle is
where position and linear momentum are polar vectors. Angular momentum is therefore axial.
Similarly, torque is
with force a polar vector. Torque is consequently a pseudovector.
Under spatial inversion, both factors in each cross product reverse:
The axial quantity remains unchanged. This statement concerns how a mirrored configuration is related to the original one; it does not imply that the angular momentum of a system is constant in time.
Angular velocity is also axial. Its direction encodes the plane and sense of rotation rather than the direction of a translational displacement.
Magnetic field
The magnetic field is an axial vector field, whereas the electric field is polar. Their different transformation laws are consistent with the Lorentz force:
Electric charge is a spatial scalar, and the cross product of the polar velocity with the axial magnetic field is polar, like force.
For a field, the transformation of its argument must also be included. Under an active orthogonal transformation,
In particular, under inversion,
The component rule does not imply that the original field has equal values at opposite spatial points.
Vorticity
For a polar velocity field , vorticity is
It is axial and describes local rotational motion in a fluid. More generally, taking the curl of a polar field produces an axial field, while taking the curl of an axial field produces a polar field.
Algebra of polar and axial quantities
A pseudoscalar is unchanged by proper rotations but changes sign under orientation-reversing orthogonal transformations. Scalar and vector products follow corresponding transformation rules:
| Operation | Transformation type of result |
|---|---|
| Polar vector polar vector | Scalar |
| Axial vector axial vector | Scalar |
| Polar vector axial vector | Pseudoscalar |
| Polar vector polar vector | Axial vector |
| Polar vector axial vector | Polar vector |
| Axial vector axial vector | Axial vector |
Here the dot denotes the Euclidean dot product. For example, if are polar vectors, the scalar triple product
is a pseudoscalar. It measures signed volume and reverses sign under reflection.
A polar vector and an axial vector cannot be added to form a quantity obeying either standard transformation law under all orthogonal transformations. Such an addition can be performed componentwise in a chosen frame, but the result is not a single polar or axial geometric quantity.
Geometric interpretation
Bivectors and the Hodge dual
A more intrinsic description of a cross product begins with exterior algebra. Given vectors and , their exterior product
is a bivector: an oriented plane element whose magnitude represents area. Unlike the three-dimensional cross product, the exterior product is defined in arbitrary dimensions without choosing a preferred handedness.
In an oriented three-dimensional Euclidean space, the metric and the Hodge star allow a bivector to be represented by a perpendicular vector. With vectors and covectors identified using the metric, this construction is commonly written
The Hodge star depends on orientation. Its behavior under orientation reversal accounts for the extra sign in the pseudovector transformation law.
This explains both the usefulness and the dimensional specificity of axial vectors. In a space of dimension , the space of bivectors has dimension
For , this equals three, allowing every bivector to be encoded by an axial vector. In four dimensions there are six independent bivector components, so the same three-dimensional representation is unavailable.
Antisymmetric tensors
A three-dimensional antisymmetric rank-two tensor has three independent components and can likewise be represented by an axial vector. In a chosen oriented orthonormal basis,
where is the Levi-Civita symbol and repeated indices are summed according to the Einstein summation convention.
The determinant factor follows from the orientation dependence of the Levi-Civita construction. Angular momentum and infinitesimal rotations can therefore be described either using axial vectors in three dimensions or using antisymmetric tensors in a form that generalizes to other dimensions.
Scope and coordinate dependence
The law
is specifically an orthogonal transformation law. It should not be applied unchanged to arbitrary coordinate transformations.
For example, if is any invertible linear transformation and the cross products are evaluated using a fixed Euclidean metric and orientation, then
Only when is orthogonal does , recovering the axial-vector rule.
In general coordinates, metric factors and the distinction between vectors, covectors, antisymmetric tensors, and tensor densities must be retained. The bivector description provides the underlying geometric object without relying on a three-component axial representation.
In representation theory, polar and axial vectors give different representations of the orthogonal group :
Their restrictions to the rotation group coincide. This is the precise algebraic reason that reflections, rather than rotations alone, are required to distinguish them.