Lorentz transformation ($\Lambda$)

A Lorentz transformation preserves the Minkowski metric. For a Lorentz boost, let $S^{\prime }$ move at constant velocity $+v$ along the $x$-axis of $S$, with parallel spatial axes and coincident origins at $t=t^{\prime }=0$. With $\beta =v/c$ and the Lorentz factor $\gamma$, the coordinates of the same spacetime event obey

The inverse Lorentz transformation replaces $v$ by $-v$. In particular, $x^{\prime }=0$ traces the moving origin $x=vt$, fixing the sign convention. The time formula is $t^{\prime }=\gamma (t-vx/c^2)$; dropping its position term would lose the relativity of simultaneity. ^[ref].

For the moving-clock example in time dilation, the second tick has $(t,x)=(5\mu \mathrm{s},3c\mu \mathrm{s})$. At $v=0.6c$, substitution gives $t^{\prime }=1.25(5-0.6\times 3)\mu \mathrm{s}=4\mu \mathrm{s}$ and $x^{\prime }=1.25(3-0.6\times 5)c\mu \mathrm{s}=0$. The clock stays at its own origin, and its coordinate time there equals its proper time.

A passive Lorentz transformation changes the components used to describe the same events and fields while preserving the standard Minkowski metric matrix: ${\Lambda }^T\eta \Lambda =\eta$. This expresses the relation between standard clock-normalized inertial frames. An arbitrary coordinate relabeling remains possible, but generally changes that metric matrix. Active boost invariance of a physical system additionally requires the relevant dynamical symmetry; a preferred physical background can obstruct it even when every equation is written in tensor notation.

An index labeling components that transform under a spacetime Lorentz transformation. For example, the four components $p^{\mu }$ form a Lorentz four-vector. Repeated upper and lower indices are contracted using the Minkowski metric.

In a station frame, a target begins at $(0,a)$ with velocity $(v,0)$, and an interceptor launches from the origin at speed u greater than v. Straight-line interception requires its velocity components to be $(v,\sqrt{u^2-v^2})$, and takes station time $t=a/\sqrt{u^2-v^2}$. The target-frame coordinate interval is ${\gamma }_v(t-v\Delta x/c^2)=t/{\gamma }_v$, since $\Delta x=vt$. This differs from the interceptor's proper time $t/{\gamma }_u$. A symmetric return to the stationary launch point instead has target-frame round-trip coordinate duration $2{\gamma }_{v}t$, because its endpoint spatial separation in the station frame is zero.

Suppose standard collinear boosts have the Lorentz transformation form, but initially assign speed scales $a$ and $b$ to two successive transformations with relative velocities $v$ and $u$. Their composition has spatial and temporal diagonal factors proportional to $1+uv/a^2$ and $1+uv/b^2$, respectively. Closure within the same calibrated transformation family requires these factors to agree, forcing $a^2=b^2$ for generic nonzero boosts.

This argument fixes a universal reciprocal squared speed; it does not establish that it is nonzero. The common limit $1/a^2=0$ gives the consistent Galilean transformation family. A finite invariant speed of light is an additional empirical property of spacetime and electromagnetism, expressed by the second of the postulates of special relativity. It cannot be deduced by excluding the Galilean limit through boost composition alone.

An orthonormal coframe in spacetime can be rotated by a position-dependent Lorentz matrix without changing the spacetime metric tensor. Its spin connection transforms as ${\omega }^{\prime }=\Lambda \omega {\Lambda }^{-1}-d\Lambda {\Lambda }^{-1}$. A spinor field transforms through a chosen spin lift ${\psi }^{\prime }=S(\Lambda )\psi$, not through the vector matrix $\Lambda$. With $S^{-1}{\gamma }^{A}S={\Lambda }^A{}_B{\gamma }^B$, an infinitesimal parameter gives $S=1+\frac{1}{4}{\alpha }_{AB}{\gamma }^{AB}+O({\alpha }^2)$, where ${\gamma }^{AB}=\frac{1}{2}[{\gamma }^A,{\gamma }^B]$. Consistent global transition lifts require a spin structure; a $2\pi$ spatial rotation lifts to minus the identity on spinors.

Spatially separated simultaneous events in one inertial frame are generally not simultaneous in a relatively moving frame. The Lorentz transformation explicitly mixes position into time. This is essential when comparing distances between moving light pulses: one must compare their positions at the same time in the measuring frame.

At $v=0.6c$, the rod's simultaneous laboratory endpoint measurements in length contraction have $(ct,x)=(0,0)$ and $(0,8\mathrm{m})$. They become $(c{t}^{\prime },x^{\prime })=(0,0)$ and $(-6\mathrm{m},10\mathrm{m})$. In contrast, simultaneous rod-frame measurements at $t^{\prime }=0$ use $(ct,x)=(0,0)$ and $(7.5\mathrm{m},12.5\mathrm{m})$. The latter front-end event occurs later in the laboratory, after the front has moved. The sloping $t^{\prime }=0$ line in the figure is a set of simultaneous spacetime events, not a signal's trajectory. ^[ref].

The apparent conflict between fitting a moving rod inside a shorter stationary enclosure in one frame and observing an even shorter moving enclosure in the rod's frame. Length contraction compares endpoints simultaneous in the measuring frame, and relativity of simultaneity means the two frames do not compare the same pairs of events. Transforming the enclosure-entry and enclosure-exit events resolves the apparent contradiction.

A runner enters a stationary tunnel of length $L$ at the tunnel-frame instant a photon is emitted towards them from the far end. The interception fraction is $f=\beta /(1+\beta )$, where $\beta =v/c$. In the runner frame, with entry at the origin, emission is at $(x^{\prime },t^{\prime })=(\gamma L,-\gamma vL/c^2)$ and the entrance clock simultaneously reads $-vL/c^2$. The entrance is then ahead by $\gamma {\beta }^{2}L$, so the photon travels $\gamma L$, not the contracted tunnel length $L/\gamma$. Subtracting that initial entrance gap from the entrance's motion during the photon flight gives the same interception fraction. This is a concrete application of relativity of simultaneity, length contraction and the Lorentz transformation.

Two spacetime events are simultaneous in a specified inertial frame when their time coordinates are equal. The Lorentz transformation changes this condition for spatially separated events. A spacelike separation admits an inertial frame in which the two events are simultaneous; a nonzero null separation does not.

In Minkowski spacetime of metric signature $(-,+,+,+)$, let $u$ be a constant future-directed four-velocity with $\eta (u,u)=-c^2$. The time difference assigned to two spacetime events is $\Delta t_u=-\eta (u,Q-P)/c^2$. Thus the observer's simultaneous events form an affine hyperplane orthogonal to $u$. This is an invariant relation of the three arguments $u,P,Q$; the Minkowski metric alone does not single out one $u$. For fixed $u$, it defines an equivalence relation and the hyperplanes form a foliation. Changing $u$ changes this foliation.

In units $c=1$, consider spacetime events $O=(0,0)$, $C=(1,2)$ and $D=(1,0)$, with coordinates ordered $(t,x)$. A Lorentz boost of speed $1/2$ assigns $t_O^{\prime }=t_C^{\prime }=0$, while the original inertial frame assigns $t_C=t_D=1$. Nevertheless $O,D$ are timelike separated, so no inertial frame regards them as simultaneous. Hence the relation “simultaneous in at least one inertial frame” is not transitive. This does not contradict transitivity of Minkowski simultaneity relative to an inertial observer for one fixed observer.

A moving rod can have both ends inside a shorter shed on a single constant-time slice of the shed frame. Lorentz transformation sends the two simultaneous endpoint events to different times in the rod frame; it does not imply simultaneous containment there. Entrance and first collision can be spacelike separated and occur in reversed time order. After impact, length changes must propagate through the rod causally; a perfectly rigid instantaneous stopping model is not relativistically admissible.

For two photons emitted simultaneously a distance $d$ apart in one frame and moving in the same positive direction, the world-line intercept $x-ct$ transforms to $\gamma (1+v/c)(x-ct)$. Their equal-time separation in the boosted frame is therefore the displayed constant $d^{\prime }$. It is different from length contraction of the emitters, because the emission events are not simultaneous after the boost.

If a differential operator $L$ is a linear map and has nonzero kernel of a linear map, the equation $LA=J$ determines $A$ only up to a homogeneous solution. Lorentz covariance of $L$ and $J$ therefore does not force independently selected solutions in different inertial frames to obey a particular tensor transformation law. A covariant Green function and consistently transformed homogeneous data do provide such a choice. For the four-potential in Lorenz gauge, $\square A^{\mu }=-{\mu }_0{J}^{\mu }$ still permits $A^{\mu }\mapsto A^{\mu }+{\partial }^{\mu }\chi$ with $\square \chi =0$, a residual Lorenz gauge symmetry. Constant spatial potentials for zero sources are a concrete example.

A physical quantity is Lorentz invariant when its value is unchanged under Lorentz transformations. Tensor equations can be Lorentz covariant while their components transform. Fixed-vector gauge choices need not make covariance manifest, although physical gauge-invariant predictions retain the symmetry.

Dynamical equations whose solutions transform into solutions under a Lorentz transformation, holding fixed the theory's physical background data. A scalar matter equation constructed from the Minkowski metric has this property; introducing a preferred timelike vector field into the matter equation can destroy it.

A Lorentz boost relates inertial frames moving at a constant relative velocity without an additional spatial rotation. In units with light speed one, the displayed transformation preserves $t^2-x^2$, with Lorentz factor $\gamma =(1-v^2{)}^{-1/2}$. Applying it to a static classical field-theory soliton produces a uniformly moving solution in a Lorentz-invariant theory.

Two noncollinear Lorentz boosts produce a Lorentz transformation with both a boost and a rotation. In the SL(2,C) lift, the product of two positive Hermitian boost matrices generally ceases to be Hermitian. Its polar decomposition of an invertible complex matrix separates a positive Hermitian boost factor and a unitary rotation factor. The rotation depends on both boosts and their order, and is called the Wigner rotation.

For boost matrices $A=cI+s\mathbf{n}\cdot \mathbf{\sigma }$ and $A^{\prime }=c^{\prime }I+s^{\prime }{\mathbf{n}}^{\prime }\cdot \mathbf{\sigma }$, the Pauli matrix multiplication law gives the displayed anti-Hermitian part. If both rapidities are nonzero and the axes are noncollinear, it cannot vanish, while either lift of a pure boost is Hermitian. Thus an additional Wigner rotation is necessary. A zero boost is a degenerate exception.

Rapidity is a dimensionless parameter for a signed collinear velocity $u$ with $|u|<c$, defined using the inverse hyperbolic functions byHere $c$ is the speed of light, and $\tanh$ is the hyperbolic tangent. Zero velocity has zero rapidity; negative velocity has negative rapidity. Every finite real rapidity gives $|u|<c$, while $u\to \pm c$ requires $\theta \to \pm \infty$. Thus a light ray has no finite rapidity in this parametrization. The header's formula uses natural units $c=1$: its $v$ is the numerical velocity $u/c$.

For the inertial frames in the velocity-addition formula, $S^{\prime }$ moves at signed velocity $V$ along the $x$-axis of $S$. Set $\varphi =\mathrm{artanh}(V/c)$. Re-expressing the same particle in the moving inertial frame givesThe subtraction is a passive change of reference frame; the inverse adds the particle's rapidity in $S^{\prime }$ to the frame's rapidity relative to $S$. For example, $V=0.6c$ and $u^{\prime }=0.6c$ give $\varphi ={\theta }^{\prime }=\ln 2$, so $\theta =\ln 4$ and $u=c\tanh (\ln 4)=15c/17$, exactly as in the velocity-addition formula.

For a massive particle of rest mass $m>0$ moving in one spatial dimension, its energy and signed momentum areThe hyperbolic cosine is the particle's Lorentz factor, and the hyperbolic sine is $\gamma u/c$. The hyperbolic Pythagorean identity immediately gives the relativistic energy-momentum relation $E^2-c^2{p}^2=m^2{c}^4$. In the older $c=1$ notation, writing the rest mass as $M$ recovers $v=\tanh \theta$, $E=M\cosh \theta$, and $p=M\sinh \theta$. These expressions use natural units in which the factors of $c$ are numerically one; the preceding formulas restore those factors explicitly. ^[ref].

The useful simplification follows from the hyperbolic-function identity $\tanh (a+b)=(\tanh a+\tanh b)/(1+\tanh a\tanh b)$: the nonlinear velocity-addition formula becomes ordinary addition of rapidities. See iterated collinear boosts with equal rapidity for a plotted example. A common collinear Lorentz boost shifts every rapidity by the same constant, so a rapidity difference is invariant. This makes rapidity useful for relativistic factorized scattering in one spatial dimension, where two-body S-matrices depend on the difference of the incoming rapidities, with particle species fixed.

If each successive inertial frame moves at the same signed velocity $u$, with $|u|<c$, relative to the preceding frame along a common axis, the rapidity adds. The final velocity relative to the first frame is therefore the displayed expression. When $u=c/N$ and $N\ge 2$, the limit is $c\tanh 1$, since $N\mathrm{artanh}(1/N)\to 1$.

Starting from rest, compose $n$ positive collinear Lorentz boosts, each with the same finite rapidity $\alpha >0$. Each step has relative velocity $w=c\tanh \alpha$, measured in the preceding rest inertial frame. The relativistic velocity-addition formula givesThese are equal increments of rapidity, not equal increments of velocity measured in the original inertial frame. They can describe a chain of relatively moving inertial frames or successive idealized changes of a massive particle's motion; no particular acceleration schedule is assumed. ^[ref].

For a relative velocity $w=c/2$ at each step,The first four velocities are $u_0=0$, $u_1=c/2$, $u_2=4c/5$, and $u_3=13c/14$. At step eight, $u_8/c=3280/3281\approx 0.99969521$. The rapidities grow by the same amount at each step, while the successive velocity increments become smaller.

A short derivation applies the hyperbolic-function identity for $\tanh (a+b)$ to the recurrence, giving ${\theta }_{n+1}={\theta }_n+\alpha$. Substituting $\tanh x=(e^{2x}-1)/(e^{2x}+1)$ and $e^{2\alpha }=3$ gives the displayed closed expression. In particular,for every finite nonnegative integer $n$. The limit is $u_n\to c$ as $n\to \infty$, but no finite sequence reaches the speed of light. This illustrates the beauty of rapidity: unbounded additive parameters coexist with bounded physical velocities. Replacing every step by a negative rapidity gives the corresponding approach to $-c$.

An active transformation keeps the coordinate argument fixed and changes a contravariant vector field by $A^{\prime \mu }(x)={\Lambda }^{\mu }{}_{\nu }{A}^{\nu }({\Lambda }^{-1}x)$. Infinitesimally, $\delta A^{\mu }=\alpha {\omega }^{\mu }{}_{\nu }{A}^{\nu }-\alpha {\omega }^{\rho }{}_{\sigma }{x}^{\sigma }{\partial }_{\rho }{A}^{\mu }$. For a Lorentz scalar density, the spin-index term is absent, giving $\delta \mathcal{L}=-\alpha {\omega }^{\rho }{}_{\sigma }{x}^{\sigma }{\partial }_{\rho }\mathcal{L}$.

A Lorentz scalar is unchanged under a Lorentz transformation. Contracting every Lorentz index in a tensor expression produces a scalar when the constituent fields transform covariantly.

A Lorentz tensor transforms by one Lorentz matrix for each contravariant index and one inverse matrix for each covariant index. Complete contraction of all indices produces a Lorentz scalar.

The Lorentz group consists of real matrices preserving the Minkowski metric: ${\Lambda }^T\eta \Lambda =\eta$.

Let $B(x)=\left(\begin{array}{cc}\cosh x& -\sinh x\\ -\sinh x& \cosh x\end{array}\right)$, $T=\mathrm{diag}(1,-1)$ and $P=\mathrm{diag}(-1,1)$. Every real matrix preserving $\mathrm{diag}(1,-1)$ has a unique form $B(x)S$ with $S\in \{I,T,P,TP\}$. The boost subgroup is normal because $TB(x)T=PB(x)P=B(-x)$. Its four cosets form the Klein four-group, giving the displayed quotient group. In contrast, the four-element subgroup generated by $T,P$ is not normal: $B(x)TB(-x)=B(2x)T$ is not diagonal for $x\ne 0$. This distinguishes the normal identity component from a chosen set of component representatives.

For $J=\mathrm{diag}(1,-1)$, the matrices satisfying $A^{T}JA=J$ and $A_{11}>0$ form the time-orientation-preserving Lorentz group in one spatial dimension. They are exactly $A=H(t)D_{\epsilon }$, where $H(t)=\left(\begin{array}{cc}\cosh t& \sinh t\\ \sinh t& \cosh t\end{array}\right)$, $D_{\epsilon }=\mathrm{diag}(1,\epsilon )$ and $\epsilon =\pm 1$. Preservation of the Minkowski metric makes the first column a future-pointing unit timelike vector; Minkowski inner product orthogonality determines the second up to sign. The identity $D_{\epsilon }H(s)=H(\epsilon s)D_{\epsilon }$ gives the multiplication law $(t,\epsilon )(s,\eta )=(t+\epsilon s,\epsilon \eta )$ and inverse $(-\epsilon t,\epsilon )$.

For $M=\mathrm{diag}(-1,1,1,1)$,Every such matrix has the block formso the Lie algebra and the Lorentz group have dimension six. At $A\in O(1,3)$, the tangent space is $A\mathfrak{o}(1,3)$.

With signature $(-,+,+)$, two Lorentz boost generators and one spatial rotation obey $[J,K_1]=K_2$, $[J,K_2]=-K_1$ and $[K_1,K_2]=-J$. Replacing the boosts by $i{K}_1,i{K}_2$ gives the compact cyclic brackets, proving the complexified isomorphism with the SU(2) Lie algebra. The real algebra has dimension three and rank one. Its finite-dimensional complex irreducible representations have dimensions $m+1$ for nonnegative integer highest weight $m$.

The orthochronous Lorentz group consists of Lorentz transformations preserving the future sheet of the unit timelike hyperboloid. Equivalently, $A_{00}>0$. This is an open subgroup of $O(1,3)$ and hence a full-dimensional submanifold.

The proper orthochronous Lorentz group is the identity component of the Lorentz group. Its elements have determinant one and preserve the future light cone.

A Lorentz transformation with determinant $+1$ that preserves future-directed timelike four-vectors. Equivalently, it belongs to the identity component of the Lorentz group. For the Minkowski metric $(-,+,+,+)$, its matrix satisfies ${\Lambda }^T\eta \Lambda =\eta$ and ${\Lambda }^0{}_0\ge 1$. Such transformations preserve the future light cone, so the retarded Green function of the d'Alembert operator transforms as a Lorentz scalar.

Use the Lorentz spinor double cover $SL(2,\mathbb{C})\to S{O}^{+}(1,3)$. For a lift with distinct eigenvalues $\lambda ,{\lambda }^{-1}$, diagonalize and choose logarithms $z,-z$; this gives a traceless matrix logarithm. For a repeated eigenvalue, choose the sign of the lift so it is $I+N$, where $N^2=0$, and use $\log (I+N)=N$. The two lifts have the same Lorentz image. Exponentiating the differential of the covering map therefore expresses every proper orthochronous transformation as the exponential of a real element of the Lie algebra of the Lorentz group. This argument does not assert that every element of $SL(2,\mathbb{C})$ itself has a traceless logarithm.

Identify Minkowski vectors with Hermitian matrices $X=x^{0}I+x^a{\sigma }_a$. The action $X\mapsto SX{S}^{†}$ preserves $\det X=(x^0{)}^2-|\mathbf{x}{|}^2$. It gives a surjective homomorphism from $SL(2,\mathbb{C})$ to the proper orthochronous Lorentz group with kernel $\{I,-I\}$. Polar decomposition retracts the covering group onto SU(2), making it simply connected. This cover permits half-integer spin group representations that do not descend to the Lorentz group itself.

A real four-vector corresponds to a two-by-two Hermitian matrix with determinant equal to its Minkowski squared norm. The action $X\mapsto NX{N}^{†}$ for $N\in SL(2,\mathbb{C})$ preserves that determinant and the future cone. Expanding with $\frac{1}{2}\mathrm{Tr}({\bar{\sigma }}^{\mu }{\sigma }_{\nu })={\delta }^{\mu }{}_{\nu }$ yields ${\Lambda }^{\mu }{}_{\nu }(N)=\frac{1}{2}\mathrm{Tr}({\bar{\sigma }}^{\mu }N{\sigma }_{\nu }{N}^{†})$. The map has kernel $\{\pm I\}$ and image $S{O}^{+}(1,3)$.

A nonzero Hermitian rank-one positive semidefinite matrix has zero determinant and positive trace, representing a future-directed null vector. It factors as $\chi {\chi }^{†}$, with an irrelevant spinor phase. A past-directed null vector instead requires a minus sign; a positive outer product alone does not cover both time directions.

With $X_{12}=x+iy$, a spinor whose components have phases $e^{-i\varphi /2}$ and $e^{i\varphi /2}$ has $x+iy\propto e^{-i\varphi }$, so its physical azimuth is $-\varphi$. The usual positive azimuth for those phases uses $X_{12}=x-iy$. Both conventions are valid if used consistently.

The real matrix space defined by $C+C^{\ast }=\mathrm{tr}(C)I$ consists exactly of $C=aI+i\mathbf{v}\cdot \mathbf{\sigma }$, with $a$ and the three components of $\mathbf{v}$ real. The Pauli matrices have $\mathrm{tr}({\sigma }_i{\sigma }_j)=2{\delta }_{ij}$, giving the displayed real symmetric bilinear form of signature $(3,1)$. The real isomorphism $C\mapsto aI-\mathbf{v}\cdot \mathbf{\sigma }$ to Hermitian matrices preserves this form when the latter have the polarized negative-determinant form $[\mathrm{tr}(AB)-\mathrm{tr}A\mathrm{tr}B]/2$.

The Poincare group is the semidirect product of spacetime translations by the Lorentz group. It is the isometry group of Minkowski spacetime.

Invariance under spacetime translations and Lorentz transformations. A scalar action with constant coefficients and contracted Lorentz indices has this symmetry when its domain and boundary conditions also permit the transformations.

The Poincare algebra is generated by translations $P^{\mu }$ and Lorentz transformations $M^{\mu \nu }$. Its mixed commutator is $[M^{\mu \nu },P^{\rho }]=i({\eta }^{\nu \rho }{P}^{\mu }-{\eta }^{\mu \rho }{P}^{\nu })$.

The Pauli-Lubanski pseudovector is $W_{\mu }={ϵ}_{\mu \nu \rho \sigma }{M}^{\nu \rho }{P}^{\sigma }/2$, up to an overall sign convention. Its square is a Casimir of the Poincare algebra and measures spin in a massive irreducible representation.

The physical massive-particle Pauli-Lubanski pseudovector obeys $W\cdot P=0$ and $W^2=-m^2{c}^2|{\mathbf{s}}_{\mathrm{rest}}{|}^2$ for classical rest-frame spin. Orthogonality by itself does not establish this normalization. In particular the components $W=(\mathbf{s}\cdot \mathbf{p},(E/c)\mathbf{s})$ give $W^2=-m^2{c}^2|\mathbf{s}{|}^2-|\mathbf{s}\times \mathbf{p}{|}^2$; dividing by $mc|\mathbf{s}|$ is a unit spacelike normalization only when the cross product vanishes.

In $W^{\mu }=-{ϵ}^{\mu \nu \rho \sigma }{P}_{\nu }{M}_{\rho \sigma }/2$, contracting with $P_{\mu }$ gives zero because the commuting momentum products are symmetric in $\mu ,\nu$ while the alternating tensor is antisymmetric. Reordering $P$ and $M$ makes no difference because the resulting metric contractions with the alternating tensor vanish.

// Target: special-relativity.bigb

For a massive positive-energy unitary irreducible representation of the Poincaré group with signature $(+---)$, choose a rest frame. The Pauli-Lubanski pseudovector has zero time component and spatial components $m{J}_i$, up to a common sign convention. Its square is therefore $-m^2{\mathbf{J}}^2$. The SU(2) little-group representation has spin $s$ and ${\mathbf{J}}^2=s(s+1)$, proving the invariant value. The mass and spin label the massive induced representation of the Poincare double cover.

With signature $(+---)$ and ${ϵ}_{0123}=+1$, a momentum $(k,0,0,k)$ and helicity $j_3$ give lower components $W_0=k{j}_3$ and $W_3=-k{j}_3$. They obey $W_{\mu }{P}^{\mu }=k(W_0+W_3)=0$. These longitudinal identities require only the momentum and rotation-projection eigenvalues, not a statement about transverse little-group translations.

With signature $(+---)$, ${ϵ}_{0123}=+1$, $J_3=M^{12}$ and the lower-index definition of $W_{\mu }$, a rest momentum $(m,0,0,0)$ gives $W_0=0$ and $W_3=-m{J}_3$ on that momentum eigenspace. A Hermitian spin projection $j_3$ consequently gives eigenvalues $0$ and $-m{j}_3$. Raising the third index changes its sign.

The little group of a momentum $p$ is its stabilizer in the Lorentz group. Its unitary irreducible representations label the internal degrees of freedom of one-particle states on the orbit of $p$.

These stabilizer representations are used in Wigner's classification of relativistic particle states.

A particle polarization specifies a physical spin state at fixed four-momentum. Massive spin states transform under the rotation little group $SO(D-1)$; finite-spin massless states transform under its transverse rotation part $SO(D-2)$. A massive two-form field has $\left(\genfrac{}{}{0ex}{}{D-1}{2}\right)$ such states, while a two-form gauge field has $\left(\genfrac{}{}{0ex}{}{D-2}{2}\right)$. Polarization components must be counted after field equations and any gauge equivalence are imposed.

For an on-shell massive vector with $k^{\mu }=(E,|\mathbf{k}|\hat{\mathbf{k}})$ in Minkowski metric signature $+---$, the displayed vector satisfies $k\cdot {ϵ}_L=0$ and ${ϵ}_L^2=-1$. It specifies the longitudinal one of the three physical polarization states. It is not a replacement for a final polarization sum when other states are unobserved.

The stabilizer of a nonzero null four-momentum has connected group $ISO(D-2)$. For ordinary finite-polarization particles its translation part acts trivially, leaving a representation of $SO(D-2)$.

For a future null standard momentum $k$, choose Lorentz transformations $L(p)k=p$ and a unitary irreducible representation $d$ of its little group. The Wigner rotation $R(\Lambda ,p)=L(\Lambda p{)}^{-1}\Lambda L(p)$ acts on the internal state. The resulting induced representation acts on wavefunctions with invariant measure $d^{3}p/(2p^0)$. Trivial little-group translations give fixed helicity; nontrivial translations give a continuous-spin representation with infinitely many polarization states.

In a finite-dimensional unitary representation of $ISO(2)$, its two commuting Hermitian translation generators have a finite joint spectrum. Conjugating by every planar rotation rotates that spectrum. A finite set invariant under all rotations contains only the origin, so both generators vanish. An irreducible representation is then a one-dimensional rotation character, giving a helicity state. Nontrivial translations instead require the infinite-dimensional continuous-spin representation. The same argument applies to the physical double cover of this little group.

For a standard null momentum $(E,0,0,E)$ in signature $(+---)$, define $K_i=M^{0i}$ and $J_i={ϵ}_{ijk}{M}^{jk}/2$. The stabilizing generators are $J_3,N_1,N_2$. The Poincare algebra gives $[J_3,N_1]=i{N}_2$, $[J_3,N_2]=-i{N}_1$ and $[N_1,N_2]=0$, the Lie algebra of planar Euclidean motions. Evaluating the Pauli-Lubanski pseudovector gives $W^2=-E^2(N_1^2+N_2^2)$.

An ordinary massless spin-one particle has transverse vector polarizations under $SO(D-2)$. A covariant polarization satisfies $p\cdot ϵ=0$ and is identified under $ϵ\mapsto ϵ+\lambda p$. The quotient has $D-2$ positive polarizations.

The stabilizer of a timelike four-momentum in $D$-dimensional Minkowski spacetime has connected group $SO(D-1)$. Massive particle polarizations form representations of this group.

A massive spin-one particle transforms as a vector under its massive particle little group $SO(D-1)$. Its $D-1$ polarizations are orthogonal to its timelike four-momentum. Under transverse $SO(D-2)$ rotations they split into a vector and a scalar.

The symmetry used to classify an open string must preserve its endpoint boundary background as well as its momentum. A Neumann-Dirichlet open-string boundary condition breaks target Lorentz transformations mixing ND and common directions with Neumann boundary conditions. For a D1–D25 pair with 24 ND directions, the common worldvolume has Lorentz group $SO(1,1)$, whose connected massive little group is trivial. The transverse $SO(24)$ remains as an internal rotation group. Thus a massive 24-component oscillator multiplet is not an $SO(25)$ vector of the full 26-dimensional massive little group; a vector of that full group would have 25 components.

A massive vector in $D$ spacetime dimensions has $D-1$ polarizations perpendicular to its timelike momentum. A massless vector has transverse polarizations modulo its null longitudinal polarization, leaving $D-2$ physical degrees of freedom.

Helicity is the projection of angular momentum onto the momentum direction. For a finite-helicity massless particle, the Pauli-Lubanski vector satisfies $W^{\mu }=\lambda P^{\mu }$.

A physical photon has two transverse helicity states. Circular polarization distinguishes their populations; a real photon has no longitudinal helicity-zero polarization.

For a relativistic neutrino, the normalized helicity is approximately $-1$ in ordinary weak production. For a massive particle, helicity and fermion chirality differ; suppressed opposite-helicity amplitudes and frame dependence prevent promoting the massless selection rule to an exact statement about arbitrary boosts.

For $\mathbf{k}=k\hat{z}$, these complex vectors form an orthonormal basis of the transverse plane using the Hermitian inner product. They have opposite phases under rotations about $\hat{z}$, so they describe the two helicity components. Since $\mathbf{e}\cdot m^{(\pm )}=\sin \theta e^{\pm i\phi }/\sqrt{2}$, transverse-vector sources carry only the angular sectors $m=\pm 1$ of spherical harmonics. Streaming by $ik\cos \theta$ preserves $m$ and couples neighboring angular degrees.

A continuous-spin representation is a massless unitary irreducible representation of the Poincare group in which the translation subgroup of the $ISO(2)$ little group acts nontrivially. It contains infinitely many helicities.

Wigner's classification obtains irreducible positive-energy representations of the Poincaré group by inducing from unitary irreducible representations of the little group of a standard momentum.

Fix $m>0$, $k=(m,0)$, and a rotationless Lorentz boost $L(p)$ taking $k$ to $p$ on the positive-energy mass shell. The little group on the Lorentz spinor double cover is SU(2). Its spin-$s$ irreducible representation induces a unitary representation on $L^2(d^{3}p/(2p^0))\otimes {\mathbb{C}}^{2s+1}$. Here $W(\Lambda ,p)=L(\Lambda p{)}^{-1}\Lambda L(p)$ is the Wigner rotation, lifted to SU(2) when necessary, and the translation convention is $U(I,a)|p⟩=e^{ip\cdot a}|p⟩$. The little-group cocycle proves the group law, and Wigner's classification gives irreducibility. Every nonnegative integer or half-integer spin is allowed on the double cover; descent to the ordinary proper orthochronous Poincaré group requires integer spin, because its $2\pi$ rotation must act as the identity.

The determinant-one Lorentz group in one space and one time dimension has two connected components. Its identity component consists of boostsand the other component is its negative.

A boost of rapidity $\phi$ acts on null coordinates by $u\mapsto e^{\phi }u$ and $v\mapsto e^{-\phi }v$, preserving $uv=x^2-y^2$.

In the $(t,z)$ plane, $u=t-z$ is retarded time and $v=t+z$ is advanced time. The radial part of the Minkowski metric is $-d{t}^2+d{z}^2=-dudv$.

For $u=t-z$ and $v=t+z$, the four-dimensional Minkowski wave operator isEvery profile depending only on one null coordinate solves the wave equation when it is uniform in the transverse coordinates.

The nonrelativistic limit takes characteristic speeds much smaller than $c$. Lorentz transformations then reduce to Galilean transformations at leading order.

The velocity-addition formula translates a particle's velocity between two specified inertial frames. Let $S^{\prime }$ move at signed velocity $V$ along the $x$-axis of $S$, with parallel axes and $|V|<c$, where $c$ is the speed of light. Define $u_x=dx/dt$ and $u_x^{\prime }=d{x}^{\prime }/d{t}^{\prime }$ for the same particle's world line, using position and time measured within the corresponding inertial frame. With Lorentz factor ${\gamma }_V=(1-V^2/c^2{)}^{-1/2}$, the forward Lorentz boost gives

For collinear motion, write $u=u_x$ and $u^{\prime }=u_x^{\prime }$. The inverse velocity-addition formula isThis is the composition of the particle's velocity in $S^{\prime }$ and that frame's velocity relative to $S$. The velocities must belong to this same chain of reference frames, with directions specified. Two arbitrary quoted speeds cannot be inserted blindly. Each velocity measurement uses distances and clocks in a single inertial frame, with Einstein synchronisation for separated clocks.

For example, let $V=0.6c$ and let the particle travel at $u^{\prime }=0.6c$ in $S^{\prime }$. Its velocity in $S$ isThe Galilean transformation would instead give $u=u^{\prime }+V=1.2c$, which fails here. For light moving along the positive axis, $u^{\prime }=c$ gives $u=c$; $u^{\prime }=-c$ similarly gives $u=-c$. Negative directions also work for massive particles: at $V=0.6c$, $u^{\prime }=-0.6c$ gives $u=0$, while $u^{\prime }=-0.8c$ gives $u=-5c/13$. Thus “addition” means composition of signed velocities, rather than addition of their magnitudes.

A short derivation differentiates the Lorentz transformation: $d{x}^{\prime }={\gamma }_V(dx-Vdt)$ and $d{t}^{\prime }={\gamma }_V(dt-Vdx/c^2)$. Dividing these differentials gives the longitudinal velocity-addition formula. The transverse formulas use $d{y}^{\prime }=dy$ and $d{z}^{\prime }=dz$, but the same transformed time denominator; omitting it would ignore the relativity of simultaneity. Replacing $V$ by $-V$ gives the inverse. When both relative and particle velocities are small compared with the speed of light, the denominator is close to one and the nonrelativistic limit recovers the Galilean transformation.

The formula lets us compare particle motion in a moving apparatus and a laboratory without losing the invariant speed of light. Einstein's 1905 paper derives the coordinate change in §3 and the composition of velocities in §5, including the collinear formula, its light-speed case, and composition of parallel transformations. ^[ref]. The rapidity parametrization makes the collinear composition particularly simple.

For future-directed unit timelike four-velocities in metric signature $(+---)$, $U_1\cdot U_2/c^2$ is the Lorentz factor of either observer measured in the rest frame of the other. In a common inertial frame it equals ${\gamma }_1{\gamma }_2(1-{\mathbf{\beta }}_1\cdot {\mathbf{\beta }}_2)$. ThereforeExpanding the numerator converts this to the dot-and-cross-product form of relativistic velocity addition. The invariant construction explains why the relative speed is symmetric even when the spatial velocity components belong to different rest frames.