{
  "id": "minkowski-space",
  "title": "Minkowski Space",
  "lang": "en",
  "source_lang": "en",
  "summary": "Minkowski space is the flat, four-dimensional spacetime of special relativity, whose metric unifies spatial distance, elapsed time, and causal structure.",
  "category": "science",
  "subcategory": "physics",
  "type": "concept",
  "aliases": [
    "Minkowski spacetime",
    "Minkowski space-time"
  ],
  "generated_at": "2026-11-03",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/minkowski-space/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/minkowski-space/",
      "json": "https://aiwiki.page/zh/minkowski-space.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Minkowski space is the mathematical model of spacetime underlying special relativity. In its standard physical form, it combines three spatial dimensions and one temporal dimension into a four-dimensional continuum equipped with a metric that distinguishes time from space. Unlike ordinary Euclidean distance, its invariant interval can be positive, negative, or zero for distinct events. This structure determines which events can influence one another, how clocks compare along different paths, and how physical quantities transform between observers.\n\nNamed after Hermann Minkowski, who developed its geometric interpretation in 1907–1908, Minkowski space provides the natural setting for relativistic mechanics, electromagnetism, and much of quantum field theory. It is also the local model for curved spacetime in general relativity. Its defining property is flatness, not the absence of acceleration, particles, or electromagnetic fields: accelerated observers and complicated physical processes can exist within an unchanged Minkowski background."
    },
    {
      "heading": "Historical development and geometric interpretation",
      "text": "Minkowski’s formulation emerged from nineteenth-century investigations of electrodynamics and the motion of bodies relative to light. James Clerk Maxwell’s electromagnetic theory identified a characteristic propagation speed, but its equations did not retain their form under the transformations of Newtonian mechanics. Hendrik Lorentz developed transformations connecting electromagnetic descriptions in relatively moving systems. Henri Poincaré subsequently clarified their mathematical structure, including their group property, and recognized the invariance of a quadratic expression involving three spatial coordinates and time.\n\nIn 1905, Albert Einstein derived relativistic kinematics from the equivalence of inertial frames and the constancy of the speed of light in vacuum. His treatment explained the relativity of simultaneity, moving-clock effects, and the transformation of lengths without requiring a mechanically specified luminiferous ether. Time coordinates were established operationally through clocks and light signals. The resulting theory already contained the physical relationships that Minkowski would reorganize geometrically.\n\nMinkowski presented an early four-dimensional formulation to the Göttingen Mathematical Society in November 1907. His best-known exposition, “Space and Time,” was delivered on 21 September 1908 at the meeting of German natural scientists and physicians in Cologne. Its central move was to treat an event as a point in a four-dimensional world and a particle’s history as a curve in that world. Separate descriptions of space and time became observer-dependent decompositions of a common geometric structure.\n\nThis approach made the Lorentz transformations analogous, though not identical, to rotations. Ordinary rotations preserve a positive-definite sum of squares; Lorentz transformations preserve a quadratic form with one sign different from the other three. Early presentations sometimes introduced an imaginary time coordinate, such as \\(ict\\), to make the formula resemble a Euclidean sum of squares. Modern treatments generally retain real coordinates and place the sign difference directly in the metric.\n\nMinkowski’s contribution was therefore more than a change of notation. It identified geometric objects—events, vectors, worldlines, and invariant intervals—that exist independently of a particular observer’s coordinates. This language helped make relativistic physical laws systematic and supplied an essential precursor to Einstein’s gravitational theory. General relativity subsequently retained the local Minkowskian structure while allowing spacetime geometry to vary from place to place."
    },
    {
      "heading": "Affine structure and the Minkowski metric",
      "text": "Strictly speaking, physical Minkowski spacetime is an affine space modeled on a four-dimensional real vector space. An affine space has points and displacement vectors but no preferred origin. Events are points; the difference between two events is a vector. Coordinates can assign four numbers to each event after an origin and a basis have been chosen, but neither choice is physically privileged. Mathematical usage also calls the underlying vector space, equipped with its indefinite bilinear form, “Minkowski space.”\n\nUsing coordinates \\(x^\\mu=(ct,x,y,z)\\), a common convention writes the metric as\n\n\\[\\eta_{\\mu\\nu}=\\operatorname{diag}(1,-1,-1,-1).\\]\n\nGreek indices run from \\(0\\) to \\(3\\), and repeated upper and lower indices are summed. For a displacement \\(\\Delta x^\\mu\\), the squared interval is\n\n\\[\\Delta s^2 =\\eta_{\\mu\\nu}\\Delta x^\\mu\\Delta x^\\nu =c^2\\Delta t^2-\\Delta x^2-\\Delta y^2-\\Delta z^2.\\]\n\nThe alternative convention \\(\\operatorname{diag}(-1,1,1,1)\\) is equally standard. It reverses the signs of squared norms without changing physical predictions, provided all formulas are adjusted consistently. The metric’s signature—the number of positive and negative directions—is its essential distinguishing feature.\n\nThe metric defines a scalar product,\n\n\\[u\\cdot v=\\eta_{\\mu\\nu}u^\\mu v^\\nu,\\]\n\nbut not a positive-definite inner product. A nonzero vector can have zero squared norm, and orthogonality behaves differently from its Euclidean counterpart. Consequently, \\(\\sqrt{u\\cdot u}\\) is not a real-valued norm on all vectors, and the Minkowski interval does not define an ordinary distance function satisfying the usual axioms of metric-space theory. “Metric” here means a nondegenerate symmetric tensor, not necessarily a positive distance.\n\nThe same tensor converts contravariant components into covariant ones:\n\n\\[v_\\mu=\\eta_{\\mu\\nu}v^\\nu.\\]\n\nWith the chosen signature, \\(v_0=v^0\\), whereas \\(v_i=-v^i\\) for spatial indices. Raising and lowering indices is thus substantive rather than decorative. It ensures that contractions, field equations, and transformation laws are expressed independently of coordinates.\n\nMore generally, an \\((n+1)\\)-dimensional Minkowski space has one temporal direction and \\(n\\) spatial directions. The physically standard case is \\(n=3\\); lower-dimensional versions are widely used in exact calculations and illustrative models. This relativistic meaning should be distinguished from other mathematical uses of Minkowski’s name, including normed-space geometry and the geometry of numbers, which concern different structures."
    },
    {
      "heading": "Light cones and causal order",
      "text": "The sign of the interval divides event separations into three invariant classes. With signature \\((+---)\\), a displacement is timelike if \\(\\Delta s^2>0\\), null or lightlike if \\(\\Delta s^2=0\\), and spacelike if \\(\\Delta s^2<0\\). These categories do not depend on the inertial observer. They express whether two events can lie on the history of a massive particle, be connected by a light ray in vacuum, or be separated too widely in space for a light signal to travel between them in the available time.\n\nAt each event, the null directions form a light cone. In standard coordinates its equation is\n\n\\[c^2\\Delta t^2=\\Delta x^2+\\Delta y^2+\\Delta z^2.\\]\n\nA choice of time orientation distinguishes the future cone from the past cone. The cone’s interior contains timelike directions, while its exterior contains spacelike directions. Diagrams often suppress two spatial coordinates and draw the remaining null rays at \\(45^\\circ\\), which assumes that the vertical axis is \\(ct\\) and that both axes have the same scale.\n\nFor timelike-separated events, an inertial frame exists in which both occur at the same spatial position. Their temporal order is the same in every inertial frame related by a future-preserving Lorentz transformation. For spacelike-separated events, an inertial frame exists in which they are simultaneous, and other frames can assign either temporal order. This reversal does not reverse a permissible signal or cause: no causal trajectory with speed at most \\(c\\) joins the two events.\n\nA nonzero null displacement has a distinctive property. Its interval vanishes even though its endpoints are different. Light emitted at one event and received at another therefore follows a path of zero proper time. There is nevertheless no inertial rest frame for light. The boost required to make a light ray stationary would have relative speed \\(c\\), outside the allowed range for transformations between massive observers.\n\nCausal structure is also deeper than a particular coordinate diagram. Multiplying the metric by a positive function preserves its null cones, although it generally changes lengths, proper times, and curvature. The cones therefore determine a conformal structure rather than a complete metric scale. Conversely, ordinary Minkowski space with its usual global time orientation has no closed future-directed causal curves: time increases along every nontrivial future-directed causal trajectory. Its standard topology and causal structure are important parts of the model, not consequences of the interval formula considered without global assumptions."
    },
    {
      "heading": "Lorentz transformations and spacetime symmetries",
      "text": "A Lorentz transformation is a linear transformation preserving the Minkowski metric. In matrix notation,\n\n\\[\\Lambda^{\\mathsf T}\\eta\\Lambda=\\eta.\\]\n\nThese transformations form the Lorentz group, denoted \\(O(1,3)\\) for the signature used here. They include spatial rotations, boosts between relatively moving inertial frames, and discrete operations such as spatial inversion and time reversal. The proper orthochronous subgroup preserves both spacetime orientation and the chosen direction of time and is continuously connected to the identity.\n\nFor a boost along the \\(x\\)-axis,\n\n\\[ct'=\\gamma(ct-\\beta x),\\qquad x'=\\gamma(x-\\beta ct),\\qquad y'=y,\\qquad z'=z,\\]\n\nwhere\n\n\\[\\beta=\\frac{v}{c}, \\qquad \\gamma=\\frac{1}{\\sqrt{1-\\beta^2}}.\\]\n\nThe mixing of \\(ct\\) and \\(x\\) expresses the fact that different inertial observers use different simultaneity hyperplanes. Setting \\(t'\\) constant does not generally set \\(t\\) constant. Space and time coordinates separately depend on the observer, while their Minkowski combination remains unchanged.\n\nBoosts become particularly transparent when expressed through rapidity, a parameter \\(\\phi\\) defined by\n\n\\[\\tanh\\phi=\\beta,\\qquad \\cosh\\phi=\\gamma,\\qquad \\sinh\\phi=\\gamma\\beta.\\]\n\nA boost is then a hyperbolic rotation in a time–space plane. Collinear rapidities add, yielding the relativistic velocity-composition rule\n\n\\[v_{\\mathrm{combined}} =\\frac{v_1+v_2}{1+v_1v_2/c^2}.\\]\n\nUnlike ordinary angular rotations, boosts follow hyperbolae of constant interval. Successive non-collinear boosts also produce a spatial rotation, an effect associated with Thomas–Wigner rotation.\n\nIncluding translations gives the Poincaré group, whose transformations have the form\n\n\\[x'^\\mu=\\Lambda^\\mu{}_\\nu x^\\nu+a^\\mu.\\]\n\nIn four dimensions it has ten continuous parameters: three rotations, three boosts, and four translations. The absence of a preferred event, spatial direction, or inertial velocity is encoded in this symmetry group.\n\nFor physical theories formulated on Minkowski space, these symmetries organize conservation laws. Under the appropriate conditions, Noether’s theorem associates spacetime translations with conservation of energy and momentum, and Lorentz symmetry with conservation of relativistic angular momentum, including the boost-related components. Lorentz covariance alone does not determine a theory’s dynamics: many different field equations share the same spacetime symmetry. It constrains how equations and quantities transform, rather than specifying every interaction."
    },
    {
      "heading": "Worldlines, proper time, and relativistic measurement",
      "text": "A particle’s history is represented by a worldline. For a massive particle, the physically allowed tangent is timelike and future-directed. The time measured by an ideal clock moving along a timelike worldline is its proper time, defined by\n\n\\[d\\tau=\\frac{1}{c}\\sqrt{\\eta_{\\mu\\nu}\\,dx^\\mu dx^\\nu}.\\]\n\nIn an inertial coordinate system this becomes\n\n\\[d\\tau=dt\\sqrt{1-\\frac{v^2}{c^2}},\\]\n\nwhere \\(v\\) is the particle’s instantaneous coordinate speed. Proper time is a property of the entire path between events, not merely of the coordinate-time difference between the endpoints.\n\nThis distinction explains time dilation. Two successive ticks of a clock at rest relative to itself occur at one spatial position in that clock’s inertial frame. Another inertial frame generally assigns a longer coordinate-time interval to those same ticks. For example, at speed \\(0.8c\\), the Lorentz factor is \\(5/3\\), so three seconds of elapsed proper time correspond to five seconds of coordinate time in that frame. Reciprocal moving-clock comparisons do not contradict one another because the observers use different sets of simultaneous distant events.\n\nThe twin paradox concerns clocks that separate and later reunite. Their elapsed times can then be compared at a common event without a distant-simultaneity convention. For fixed timelike-separated endpoints in Minkowski space, the straight inertial worldline has the greatest proper time among future-directed timelike paths joining them. A traveling clock that changes velocity to return generally accumulates less proper time. Acceleration is necessary to alter its inertial path in the usual setup, but the elapsed time is calculated from the worldline integral, not from an extra correction proportional to acceleration.\n\nFour-dimensional kinematics packages motion into the four-velocity,\n\n\\[U^\\mu=\\frac{dx^\\mu}{d\\tau}, \\qquad U\\cdot U=c^2,\\]\n\nand four-acceleration,\n\n\\[A^\\mu=\\frac{dU^\\mu}{d\\tau}.\\]\n\nDifferentiating the normalization gives \\(U\\cdot A=0\\). Acceleration is therefore orthogonal to velocity in the Minkowski sense, even though this relationship has no direct analogue for ordinary three-velocity and three-acceleration.\n\nSpatial measurement also requires a simultaneity prescription. A rod’s length in a given inertial frame is determined by recording both endpoints at the same time in that frame. Lorentz contraction compares such frame-dependent measurements, rather than comparing two endpoints selected independently of time. More generally, an observer with four-velocity \\(U\\) identifies instantaneous spatial displacements with the subspace orthogonal to \\(U\\). The metric restricted to that subspace, with its sign reversed, is positive definite and supplies the observer’s local Euclidean spatial geometry."
    },
    {
      "heading": "Relativistic mechanics and classical fields",
      "text": "Minkowski geometry unifies mechanical energy and momentum in the four-momentum. For a particle of constant rest mass \\(m\\),\n\n\\[p^\\mu=mU^\\mu =\\left(\\frac{E}{c},\\mathbf p\\right).\\]\n\nIts invariant norm gives\n\n\\[p_\\mu p^\\mu=m^2c^2, \\qquad E^2=\\mathbf p^2c^2+m^2c^4.\\]\n\nThe rest-frame value \\(E=mc^2\\) is the familiar mass–energy relation. For a massless particle, the four-momentum is null and \\(E=|\\mathbf p|c\\). The geometrical distinction between timelike and null vectors thus has a direct dynamical expression.\n\nFour-momentum conservation provides an efficient method for analyzing collisions and decays. The sum of incoming four-momenta equals the sum of outgoing four-momenta in every inertial frame. The invariant mass of a collection of particles depends on its total four-momentum and need not equal the sum of the particles’ individual rest masses. A pair of oppositely directed photons, for example, can have zero total spatial momentum but positive total energy and therefore a positive invariant mass.\n\nFor a free massive point particle, the relativistic action can be written\n\n\\[S=-mc^2\\int d\\tau.\\]\n\nIts stationary paths are straight timelike worldlines when expressed in inertial coordinates. This formulation connects the geometrical description to variational mechanics. External forces change the worldline without requiring spacetime itself to become curved; curvature of a trajectory and curvature of the ambient spacetime are distinct concepts.\n\nElectromagnetism is especially natural in this framework. The electric field and magnetic field combine into the antisymmetric electromagnetic field tensor \\(F_{\\mu\\nu}\\), while charge density and current combine into a four-current. Different observers decompose the same tensor into different electric and magnetic parts. A field that is purely electric in one frame can therefore have both electric and magnetic components in another. This is a transformation of the field’s measured components, not a change in the underlying physical field.\n\nMaxwell’s equations become tensor equations on Minkowski spacetime, and the Lorentz force becomes a covariant equation relating four-momentum change to the electromagnetic tensor and four-velocity. Continuous matter and fields are similarly described by a stress–energy tensor \\(T^{\\mu\\nu}\\), which contains energy density, momentum density, and stresses. For an isolated total system in Cartesian inertial coordinates,\n\n\\[\\partial_\\mu T^{\\mu\\nu}=0.\\]\n\nWhen matter exchanges energy and momentum with an electromagnetic field, the matter and field tensors need not be separately conserved; their combined tensor obeys the conservation equation."
    },
    {
      "heading": "Accelerated coordinates and flat-space geometry",
      "text": "Minkowski space does not require every observer or coordinate system to be inertial. An accelerated observer can use coordinates adapted to a changing state of motion, and the metric components in those coordinates can depend on position or time. Such dependence does not by itself indicate gravitational curvature. The coordinate-independent test is the Riemann curvature tensor, which vanishes everywhere in Minkowski space.\n\nEven ordinary spherical spatial coordinates produce a nonconstant metric:\n\n\\[ds^2=c^2dt^2-dr^2-r^2d\\theta^2 -r^2\\sin^2\\theta\\,d\\varphi^2.\\]\n\nThe factors of \\(r\\) and \\(\\sin\\theta\\) describe the geometry of the coordinates. They do not make spacetime curved. Likewise, the coefficients appearing in the coordinate form of the geodesic equation can be nonzero in flat spacetime; they vanish globally only in suitable affine inertial coordinates.\n\nUniform linear acceleration supplies a more specifically relativistic example. A worldline with constant proper acceleration \\(a\\) can be parameterized as\n\n\\[ct(\\tau)=\\frac{c^2}{a}\\sinh\\left(\\frac{a\\tau}{c}\\right), \\qquad x(\\tau)=\\frac{c^2}{a}\\cosh\\left(\\frac{a\\tau}{c}\\right).\\]\n\nIt lies on the hyperbola \\(x^2-c^2t^2=c^4/a^2\\). Its coordinate speed approaches \\(c\\) at large positive inertial time, while its locally measured proper acceleration remains constant. The limiting speed is never reached at a finite proper time.\n\nRindler coordinates describe a family of such accelerated observers in the right-hand wedge \\(x>|ct|\\). Defining\n\n\\[ct=\\rho\\sinh\\left(\\frac{a\\eta}{c}\\right), \\qquad x=\\rho\\cosh\\left(\\frac{a\\eta}{c}\\right),\\]\n\ngives\n\n\\[ds^2=\\left(\\frac{a\\rho}{c}\\right)^2d\\eta^2 -d\\rho^2-dy^2-dz^2.\\]\n\nHere \\(\\rho>0\\), and an observer at \\(\\rho=c^2/a\\) has proper time \\(\\eta\\). Observers at other fixed values of \\(\\rho\\) have different proper accelerations, \\(c^2/\\rho\\). This variation is required for the congruence to maintain fixed neighboring proper separations.\n\nThe wedge boundaries are acceleration horizons: the eternally accelerated observers do not have access to all of Minkowski spacetime by exchanged signals. These horizons do not mark curvature singularities or physical walls. They arise from the restricted worldlines and coordinate domain.\n\nRindler time also illustrates why coordinate clocks at different positions can run at different rates in flat spacetime. This effect resembles gravitational clock comparisons in an accelerated laboratory and is closely related to the equivalence principle. It does not imply that every gravitational field is merely a coordinate effect; tidal curvature cannot be removed by an accelerated change of coordinates."
    },
    {
      "heading": "Relation to curved spacetime and global geometry",
      "text": "General relativity replaces the globally fixed Minkowski metric with a spacetime-dependent Lorentzian metric \\(g_{\\mu\\nu}\\). At each event, its tangent space has Minkowski signature. Coordinates can be chosen near that event so that\n\n\\[g_{\\mu\\nu}(p)=\\eta_{\\mu\\nu}, \\qquad \\partial_\\alpha g_{\\mu\\nu}(p)=0.\\]\n\nThese equalities express the local inertial character of spacetime. They do not generally eliminate second derivatives of the metric or the curvature built from them. A freely falling laboratory can remove the leading coordinate effects of gravity at a point, but it cannot eliminate genuine tidal effects throughout an extended region.\n\nThe distinction can be seen through geodesic motion. Freely moving test particles follow geodesics in both flat and curved spacetime. In Minkowski space, initially parallel inertial worldlines remain parallel. In curved spacetime, the separation of neighboring geodesics can change because of curvature, as described by geodesic deviation. This relative acceleration provides an operational way to distinguish gravitational tidal fields from the effects of choosing accelerated coordinates in flat spacetime.\n\nThe Einstein field equations relate spacetime curvature to stress–energy and a possible cosmological constant. Minkowski spacetime is a solution when the stress–energy tensor and cosmological constant both vanish. Vacuum alone is not sufficient to imply Minkowski geometry: gravitational waves and the exterior region of an isolated spherical mass can have nonzero curvature despite vanishing local matter stress–energy. The curvature that describes tidal gravity contains more information than the part directly fixed by the local matter tensor.\n\nMinkowski space also serves as a controlled approximation. In weak-field calculations one often writes\n\n\\[g_{\\mu\\nu}=\\eta_{\\mu\\nu}+h_{\\mu\\nu},\\]\n\nwhere the perturbation \\(h_{\\mu\\nu}\\) is small in a specified coordinate system. This is the starting point for linearized gravity and many calculations involving gravitational waves. The approximation must be judged over the spatial scales, durations, and precision relevant to a problem; a small local departure does not guarantee negligible accumulated effects over long distances.\n\nGlobal structure introduces a further distinction. Standard Minkowski spacetime has the topology of \\(\\mathbb R^4\\), admits global inertial coordinates, and possesses complete straight geodesics. A spacetime can be locally flat without being globally identical to it. Identifying one spatial coordinate periodically, for example, produces a flat spacetime with a compact spatial direction. Local measurements still give the Minkowski interval, but global paths and topology differ. Thus “flat Lorentzian spacetime” and “standard Minkowski spacetime” are not interchangeable without assumptions about topology, completeness, and identifications."
    },
    {
      "heading": "Quantum fields, particle states, and analytic continuation",
      "text": "Relativistic quantum field theory commonly begins with fields defined on Minkowski spacetime. Scalar, spinor, and tensor fields transform according to different representations of the Lorentz group, while the Poincaré group organizes the possible states of isolated particles. In Wigner’s classification, irreducible positive-energy particle representations are characterized by invariants including mass and spin or, for the familiar massless cases, helicity. The mass-shell relation is the quantum counterpart of the invariant four-momentum relation in relativistic mechanics.\n\nThe causal geometry places restrictions on observables. In a local relativistic theory, observables associated with spacelike-separated regions commute, a condition called microcausality. For fundamental fermionic fields the corresponding locality condition uses anticommutators; measurable local observables obey the appropriate commuting locality structure. This framework is compatible with quantum entanglement: nonclassical correlations between spacelike-separated measurements do not provide a controllable faster-than-light signal.\n\nMinkowski geometry also determines basic wave operators. With the metric convention used here, the d’Alembertian is\n\n\\[\\Box=\\eta^{\\mu\\nu}\\partial_\\mu\\partial_\\nu =\\frac{1}{c^2}\\frac{\\partial^2}{\\partial t^2}-\\nabla^2.\\]\n\nA free scalar field of mass \\(m\\) satisfies the Klein–Gordon equation,\n\n\\[\\left(\\Box+\\frac{m^2c^2}{\\hbar^2}\\right)\\phi=0,\\]\n\nwhere \\(\\hbar\\) is the reduced Planck constant. Retarded solutions propagate disturbances within or on the future light cone. Time-ordered quantum propagators have a different mathematical role: their nonzero values at spacelike separation do not by themselves represent superluminal transmission.\n\nThe Minkowski vacuum is invariant under the proper orthochronous Poincaré group in the standard formulation. Nevertheless, the response of a detector depends on its trajectory. The Unruh effect predicts that an ideal uniformly accelerated detector coupled appropriately to a field in the Minkowski vacuum has a thermal response with temperature\n\n\\[T_{\\mathrm U}=\\frac{\\hbar a}{2\\pi c k_{\\mathrm B}},\\]\n\nwhere \\(k_{\\mathrm B}\\) is the Boltzmann constant. This connects the accelerated-coordinate horizons of flat spacetime with quantum field behavior without requiring nonzero spacetime curvature.\n\nA separate calculational technique, Wick rotation, relates some Minkowski-signature expressions to Euclidean ones by analytic continuation of time. Formally setting \\(t=-i\\tau_{\\mathrm E}\\) turns oscillatory factors into exponentially weighted expressions under suitable conditions. This is important in perturbative calculations and lattice formulations of quantum field theory. It is not an ordinary real coordinate transformation: it changes the mathematical domain and requires control of singularities, boundary conditions, and analytic properties. Recovering real-time observables from Euclidean data is consequently a substantive mathematical problem, rather than simply relabeling an axis."
    }
  ],
  "links": [
    "spacetime",
    "special-relativity",
    "electromagnetism",
    "quantum-field-theory",
    "general-relativity",
    "james-clerk-maxwell",
    "albert-einstein",
    "speed-of-light",
    "vector-space",
    "light-cone",
    "lorentz-transformation",
    "lorentz-group",
    "noethers-theorem",
    "energy",
    "momentum",
    "worldline",
    "proper-time",
    "time-dilation",
    "twin-paradox",
    "mass-energy-equivalence",
    "action-physics",
    "electric-field",
    "magnetic-field",
    "maxwells-equations",
    "lorentz-force",
    "riemann-curvature-tensor",
    "equivalence-principle",
    "geodesic",
    "einstein-field-equations",
    "gravitational-wave",
    "quantum-entanglement",
    "planck-constant",
    "boltzmann-constant",
    "wick-rotation"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}