Classical Mechanics from Zero in Two Languages
Introduction
Everything you have ever watched move did so in a world you could point to. A thrown stone traces an arc, a spun top keeps its axis and then wanders, a door swings on its hinge and a gyroscope resists the hand that tries to tilt it. Each is a story told over time about where a body is and which way it faces. Classical mechanics makes those stories precise. Its first act, before any force is named, is to fix a stage for position and orientation. That stage is a reference frame: a choice of a point to call the origin and a set of directions against which every other direction is measured. A frame is inertial when a body left alone moves in a straight line at constant speed within it. Then we may stop distinguishing a place from the arrow that reaches it. A point of space and the vector that runs from the origin to become two names for one thing. All of mechanics can then be written in the language of vectors.
There is more than one such language, and the choice between them matters. This article develops two of them side by side and keeps them in step from the first definition to the last. The first is the matrix and linear algebra most readers know: a vector is a column of numbers, a rotation is an orthogonal matrix and an angular velocity is a vector extracted from an antisymmetric matrix by a memorised rule. The second is geometric algebra, the Clifford-algebra language in which the product of two vectors is a single object that contains both their common projection and their oriented plane. A rotation is generated by that plane directly and the pseudo-quantities of the first language become first-class elements of the algebra. The two describe the same mechanics. The matrix language keeps every formula computable and the geometric language keeps every formula meaningful.
Part I fixes the algebra of a single instant: the inertial arena and the identification , then the vector space of arrows, the inner product that measures them and the linear maps that move them, each stated first in matrix terms. We then press on the matrix language until its seams show and answer with the geometric product, the graded algebra it generates, and the reflections and rotors in that algebra. Part II treats the calculus the same way and replaces the grad, divergence and curl of vector analysis and their separate integral theorems by one vector derivative and one fundamental theorem. Part III applies the machinery to physics: Newton's laws, angular momentum as an oriented plane, rotating frames, and the rigid body whose orientation is a rotor that never locks. Part IV lifts all of this to the Lagrangian and Hamiltonian formalisms in both languages and closes by comparing them.
Three threads run the length of the article. The first is a set of discomforts of the matrix language: a rotation of the plane has no real eigenvalue, the axis of a spatial rotation is an accident of three dimensions, the cross product is a pseudovector that reflects with the wrong sign and replaces an oriented plane by a normal vector only in three dimensions, and the determinant and the notion of orientation sit outside the algebra of vectors. The second thread is the answer to each, delivered by the geometric product and the objects it makes. The third thread is a single construction, specialised once. We build the geometric algebra of a general finite-dimensional real inner-product space once and in full. Then we read off , the algebra of physical space in which the rest of the article works. The facts special to three dimensions are flagged as dimensional accidents. Chief among them are the cross product's identification with a normal vector and the bivector-to-vector Hodge dual.
The reader is assumed to have met vectors, partial derivatives and a little linear algebra: the multiplication of a matrix by a column and the idea of a basis. No prior acquaintance with geometric algebra is presumed. Nor is any with the abstract theory of manifolds: the inertial frame keeps us among ordinary vectors in . Every symbol is defined where it first appears and every step is shown.
Part I
1. Two languages
1.1. Inertial arena
Before any vector can be written the space of positions must be fixed. Physical space has no distinguished point and no distinguished directions. A stone falls the same way in Edinburgh and in Quito. A space of positions with no preferred origin is modelled by an affine space, and choosing an origin turns it into the vector space in which we compute.
An affine space modelled on a real vector space is a set of points together with a map that assigns to each ordered pair of points a vector , subject to two demands: for any point and any vector there is exactly one point with , and for any three points the displacements chain,
The space has points and their differences only. It has no origin and no way to add two points.
An affine space has differences of positions but not positions themselves. This matches the physics: only displacements, velocities and separations are measurable, never an absolute location. The physicist supplies a reference point by choice.
An inertial frame is a choice of a point , the origin, in an affine space whose modelling vector space has the inner product of Section 1.3, such that a body subject to no force has tracing a straight line at constant rate. Relative to every point acquires a unique position vector
and the assignment is a bijection from points to vectors. We write and from here on name points by their position vectors.
The identification of [eq:position-vector] is the move that spares the article the machinery of manifolds. A point is now a vector and a trajectory is now a vector-valued function . Velocity and acceleration along it are vectors in the same space as the positions. The chaining rule of [eq:affine-chain] states that displacement from to is the difference of their position vectors. This is the only sense in which positions may be subtracted.
A moving body is such a vector-valued function of time. The arena therefore already describes how it moves, before is axiomatised in Section 1.2 and long before the geometric calculus of Part II. All that is needed is the derivative of a single-parameter function, the limit of a difference quotient of vectors.
The motion of a body is a trajectory, a vector-valued function assigning to each instant its position vector . Its velocity and acceleration are the first and second time derivatives, each the limit of a difference quotient of vectors,
They are the arrows tangent to the path, vectors in the same space as the position, and the overdot abbreviates .
Velocity and acceleration lie in the same as the position. That space is the one Section 1.2 axiomatises and the length that turns a velocity into a speed is the inner product of Section 1.3. Position, velocity and acceleration at one instant form a single object, the jet of the trajectory there.
The k-jet of a trajectory at time is the tuple of its value and first derivatives there,
The 2-jet packages position, velocity and acceleration as one datum. It is the data a law of motion constrains: the second law of Section 3.2 fixes the acceleration entry from the position and closes the 2-jet into a trajectory.
These derivatives are limits and every later estimate rests on reading them as approximations with a controlled error. The derivative at an instant is the best linear approximation to the trajectory there. It is a bounded linear map whose size is an operator norm, and that number controls both the Landau remainder of a first-order expansion and the Lipschitz growth of the displacement.
The derivative is the best linear approximation to the trajectory at , the unique linear map for which the increment splits into a linear part and a remainder that dies faster than ,
As a bounded linear map its size is the operator norm , the speed itself. When the velocity is continuous on , the increment obeys the mean-value estimate
so a bound on the speed is a Lipschitz bound on the displacement.
Proof
The linear part is unique: if two linear maps both leave a remainder that is , their difference is a linear map that is itself , and a linear map vanishing faster than its argument is zero. The size of the one-input map is , the speed. For the increment, write the displacement as the integral of the velocity, , and bound the integral by the supremum of the integrand over the interval times the interval length, which is the estimate [eq:lipschitz-bound].
The estimate [eq:lipschitz-bound] is one rung of a ladder that runs the length of the jet. Each entry of is the derivative of the one before it. The same lemma applied one level up therefore reads the acceleration as the operator norm that controls the velocity: , and one level higher the jerk controls the acceleration in the same way. A bound on the size of jet level is a Lipschitz bound on the variation of level across a time step. Control enters at the top of the jet and descends one rung at a time. This is all the functional analysis the article needs: one operator-norm estimate repeated up the tower of derivatives governs every level's increment by the size of the level above it. Estimates on the acceleration therefore propagate down to the velocity and the position.
The jet is the calculus's graded object, to be set beside the algebra's. The -jet stacks a trajectory's data by order of differentiation exactly as the multivector of Section 1.7 will stack the algebra's data by grade. Each is a tower of independent levels. The correspondence is an analogy only. Raising the grade of a multivector by the outer product echoes raising the order of a jet by differentiation, and the operator norm that bounds adjacent jet levels plays the role the geometric product plays between adjacent grades. Where the algebra reads a piece off with the grade projection the calculus reads a level off with the derivative.
The operator norm of [eq:increment-bound] is the derivative's size in . The estimates therefore depend on having a length, as velocity and acceleration do. Everything rests on the vector space : its axioms, fixed in Section 1.2, define the arrows, and the inner product of Section 1.3 gives them the length that turns a velocity into a speed and the operator norm into a number. The full geometric calculus, which differentiates fields as well as trajectories, and its fundamental theorem come in Part II. We axiomatise next.
1.2. Vector space
The position vectors of Section 1.1 lie in a set with two operations, addition of vectors and scaling of a vector by a real number. This section states the axioms they obey, fixes the model and defines dimension and basis, the two structural invariants used throughout.
A real vector space is a set with an addition and a scalar multiplication such that addition is associative and commutative with an identity element and inverses, and the two operations interlock through
for all scalars and all vectors . The elements of are vectors and the elements of are scalars.
These axioms say that vectors may be added and scaled coherently. They say nothing of length or angle, which arrive in Section 1.3. A finite set of vectors is linearly independent when the only way to combine them to the zero vector is with all coefficients zero, and a basis is a linearly independent set whose linear combinations exhaust . Every basis of a given has the same number of elements, and that number is an invariant of the space.
A basis of is a set that is linearly independent and spans , so that every has a unique expansion
The number , common to all bases, is the dimension , and the real numbers are the components of in that basis. We write the component index as a superscript.
Fixing a basis turns each abstract vector into the column of its components. This is the explicit model in which the matrix language computes.
Once a basis is chosen, the map sending each vector of to its component tuple is a bijection onto that respects addition and scaling. We display the tuple as a column vector
and identify with through it. For the physics of the later parts throughout, though the algebra of this part is stated for general .
The vector is one object while the column of [eq:column-vector] depends on the basis: a different basis gives the same arrow a different column. This distinction between an object and its basis-dependent representation is the seam that Section 1.4 exploits for maps and Section 1.5 presses until the matrix language complains. still lacks any notion of length or angle. We supply it now.
1.3. Inner product
The vector space of Section 1.2 adds and scales arrows but cannot say that one is longer than another or that two meet at a right angle. Mechanics needs both: kinetic energy is built from speed and work from the angle between force and displacement. A single symmetric positive bilinear form supplies length and angle at once.
An inner product on a real vector space is a map that is
and positive-definite, meaning for every . We write for and call the pair a Euclidean space.
From the single form both metric notions follow. The norm or length of a vector is , well defined because positive-definiteness makes the radicand non-negative. The angle between two nonzero vectors is fixed by
the fraction lying in by the Cauchy-Schwarz inequality, and two vectors are orthogonal when . In components, writing for the matrix of the form in a basis, the inner product is , summed over repeated indices.
A basis is most convenient when orthonormal: the matrix is then the identity and components read off directly as inner products.
A basis of is orthonormal when
that is, each basis vector has unit length and distinct basis vectors are orthogonal. In such a basis and .
Every Euclidean space admits an orthonormal basis. The Gram-Schmidt process produces one from any basis: keep the first vector, subtract from each later vector its projections onto the directions already fixed, then normalise. We state the process and then run it in full.
Given any basis of a Euclidean space, define recursively
Then is an orthonormal basis, and each partial set spans the same subspace as .
Proof
Each of [eq:gram-schmidt] is with its components along the earlier removed, so for each , and dividing by the length makes a unit vector; induction on makes the whole set orthonormal. Because is plus a combination of the earlier vectors, with the coefficient of equal to one, the partial sets and span one subspace, and the independence of the input basis keeps every nonzero, so the division is legitimate.
Work in with the standard inner product , and orthonormalise the basis
taking one vector at a time through [eq:gram-schmidt].
Start with , whose squared length is , so .
Next remove from its component along . The overlap is , so
Its squared length is , giving after clearing the half.
Finally remove from its components along both. The overlaps are and , so
with squared length , so .
Check orthonormality of [eq:orthonormal] on the mixed pairs: , , and , while , , and are each . The three vectors form the orthonormal basis [eq:gram-schmidt] promises.
The inner product completes the Euclidean space that the whole article inhabits. It is also the seed of the second language: the geometric product of Section 1.6 is built on the demand that a vector times itself return the squared length of this section. First we give the maps between vector spaces their matrix form.
1.4. Linear maps and their matrix representation
Mechanics is the study of how vectors transform: how a rotation carries one orientation to another, how the inertia of a body turns its angular velocity into its angular momentum. The transformations that respect the vector-space structure are the linear maps, and once a basis is fixed each acquires a matrix. This section defines both and records how the matrix changes with the basis, which Section 1.5 turns into a complaint.
A linear map between real vector spaces is a function that respects addition and scaling,
When the map is a linear operator on . A linear map is determined by its values on a basis, since follows from [eq:linear-map].
A linear map is fixed by its action on a basis and each image is a vector expandable in that basis. Hence the whole map is an array of numbers.
Let be a linear operator on and a basis. Expanding each image in the basis,
defines the matrix of in that basis, whose -th column is the component tuple of . The action of on a vector becomes matrix-times-column, , and the composite of two operators has matrix the product .
The operator is one object and its matrix is a shadow cast by the basis, as the column of [eq:column-vector] is the shadow of the vector . Another basis gives the same operator a different matrix, conjugate to the first.
Let and be two bases, related by the invertible change-of-basis matrix whose columns give the new basis vectors in the old, . Then the matrices of a fixed operator in the two bases are conjugate,
Quantities unchanged by such conjugation, the determinant and the trace among them, are properties of itself.
Proof
The two columns of a single vector are related by , since the new basis vectors are . Applying this to both and its image, and using in the old basis and in the new, gives for every column , so and . The determinant is then invariant because , and the trace because it is cyclic, .
The conjugation law of [eq:conjugation] states that a matrix is a representation. It separates what belongs to the operator from what belongs to the observer's choice of frame. The determinant, trace and eigenvalues are unchanged by the change of basis and so mean something physical. The individual entries do not. The matrix language is now complete and we press on it until it strains.
1.5. Discomforts of matrix theory
The matrix language of Sections 1.1 through 1.4 is complete and computable and suffices for all of mechanics. This section lists the places where that language, though correct, describes the geometry through a veil. The geometric product and the objects it makes answer each discomfort in the same order, in Sections 1.6 through 1.11.
Rotations of the plane have no real eigenvalue. An eigenvector of an operator is a direction the operator only stretches. No notion in a matrix is more basis-independent. Yet the rotation of the Euclidean plane through an angle ,
has characteristic roots , which are not real unless is a multiple of . The most elementary rotation turns no real direction into a multiple of itself. Its eigen-analysis therefore leaves the real plane for the complex numbers. The rotation is a real motion and the appearance of is an artefact of asking a matrix for eigenvectors it does not have.
Spatial rotation axes are an accident of odd dimension. In three dimensions a rotation fixes a real direction, its axis, because a real orthogonal matrix of determinant has a characteristic polynomial of odd degree. Its non-real roots come in conjugate pairs of modulus one whose products are . The real roots are therefore each with product , and one of them is . The eigenvector for that root is the axis. The argument uses the oddness of the dimension and the sign of the determinant, and nothing special to three. In the plane there is no axis. In four dimensions a rotation can turn every direction at once with no axis anywhere. The familiar picture of a rotation as a spinning about a line is a coincidence of the number three. The plane in which the turning happens exists in every dimension and the axis perpendicular to it does not.
Cross products are three-dimensional pseudovectors that reflect with the wrong sign. The cross product is the operation used throughout mechanics to make an axis of rotation, a torque or an angular momentum. The vector it returns stands in for an oriented plane. This stand-in exists in three dimensions alone, because only there does the space of oriented planes have the same dimension, three, as the space of vectors. A plane can then be exchanged for the vector normal to it. The exchange hides a choice of handedness that appears under reflection. A proper vector such as a velocity reverses when the world is reflected in a mirror. Under the point reflection the cross product of two reflected vectors comes back unreversed, because the two sign flips cancel where a proper vector takes one. Under a mirror it picks up the factor , , and that sign is the buried handedness. An object that behaves as a vector under rotations but fails to flip under reflection is a pseudovector, and torque and angular velocity are pseudovectors. The figure shows the failure.
The plane swept by and reflects faithfully and the arrow that stands in for it flips the wrong way. The oriented plane is the invariant object. Section 1.6 names it the bivector, which needs no disguise.
Orientation and the determinant are bolted on from outside. Whether a basis is right- or left-handed and whether a linear map preserves or reverses that handedness is measured by the sign of the determinant of [eq:conjugation]. The determinant is an invariant but in the matrix language it is a separate formula, an alternating sum over permutations, imported to report on volume and orientation. It is not the product of anything within the algebra of vectors. No operation on vectors outputs the oriented volume they span. Orientation sits outside the vector calculus.
There is no closed product of vectors. One absence underlies all of the above. Real numbers may be added and multiplied and the product of two numbers is a number. Vectors may be added and scaled and fed to an inner product that returns a scalar or to a cross product that in three dimensions returns a pseudovector. There is no associative multiplication of two vectors that yields an object of the same kind and closes the operation as multiplication closes on numbers. The inner product forgets all direction and keeps a scalar. The cross product keeps a direction only by exchanging an oriented plane for its normal. That exchange is possible in three dimensions alone. Neither is a product on which an algebra can be built. Every discomfort in this list is a symptom of that lack, and the remedy is a single new product, defined next.
1.6. Geometric product
The remedy to every discomfort of Section 1.5 is a single associative product on vectors that closes: it multiplies two elements of one algebra and returns an element of the same algebra, it contains the inner product as a part and it generates the oriented planes that the cross product could only impersonate. We demand one thing of it: a vector times itself returns its squared length. Everything else in this part is read off from that demand.[1][2]
On the Euclidean space of Section 1.3 the geometric product is an associative, bilinear, unital product written by juxtaposition , subject to the single contraction axiom that every vector squares to its inner-product norm,
a scalar, for every . The algebra generated by under this product is the geometric algebra of , developed abstractly in Section 1.7 as .
The product is bilinear and associative like matrix multiplication. Commutativity is not assumed, and the geometry lies in that omission. Polarise the axiom of [eq:contraction] by applying it to a sum and expanding without reordering factors,
The left side is by the axiom and the bilinearity of , while the outer terms on the right are and . Cancelling them leaves the symmetric part of the product equal to the inner product,
The symmetric part of the geometric product is the inner product of Section 1.3. The antisymmetric part receives its own name.
For vectors the inner product and the outer product are the symmetric and antisymmetric parts of their geometric product,
so that the geometric product decomposes as
The outer product is a new kind of object, a bivector, and it is antisymmetric, , so that .
A single product of two vectors contains two pieces of information: the scalar that measures how much the vectors align and the bivector that measures the oriented plane they span. Its magnitude is the area of the parallelogram they subtend and its orientation is the sense of circulation from to . The bivector is the invariant that Section 1.5 sought in place of the cross product: an oriented plane element defined in every dimension that reflects faithfully because it is built from the vectors themselves and hides no handedness. The figure shows both parts moving together as and sweep.
In the grade language of Section 1.7 these two pieces are the grade- and grade- parts of the product, written and . Hence the decomposition of [eq:geometric-decomp] reads
This is the grade--times-grade- instance of a single associative product defined on multivectors of every grade, which Section 1.7 builds in general. The vector product here is its lowest case.
Two special cases fix the intuition. When and are parallel their outer product vanishes and the geometric product is the pure scalar . The contraction axiom of [eq:contraction] is the case . When they are orthogonal their inner product vanishes and the geometric product is the pure bivector , which then anticommutes, . Orthogonal vectors anticommute and parallel vectors commute, and the general product of [eq:geometric-decomp] interpolates between the two by the angle. The next section builds this closed associative product out to all grades.
1.7. Graded structure and the universal property
The outer product of Section 1.6 takes two vectors to a bivector and nothing stops us wedging a third. This section extends the outer product to all degrees, names the graded family it produces, builds the grade projection and the general geometric product on multivectors of every grade, and defines the geometric algebra by a universal property from which its dimension follows. The construction is done once in general dimension and specialised to physical space in Section 1.8.
The outer product extends to any number of vectors as the fully antisymmetric part of their geometric product. A product of mutually orthogonal vectors is already fully antisymmetric and is the model.
The outer product of vectors is the totally antisymmetric part of their geometric product; it vanishes when the vectors are linearly dependent and otherwise represents the oriented -dimensional volume element they span. Such an object is a blade of grade , or a -blade. Grade is the scalars, grade the vectors, grade the bivectors, grade the trivectors, and so on. A multivector is a general element of the algebra: a sum of blades of various grades. When every term of such a sum shares one grade the multivector is homogeneous of grade , a -vector; a -blade is the special -vector that factors as a single outer product of independent vectors, and a general -vector is a sum of these. The -vectors form a linear subspace of dimension , one basis blade for each choice of of the basis directions.
A multivector splits uniquely into homogeneous pieces. The algebra therefore has an operator that reads each piece off.
The grade- projection is the linear operator returning the grade- part of a multivector : it keeps the grade- blades in and annihilates every other grade. Each multivector is the sum of its projections,
with its scalar part, its vector part, and so on to the top grade . The projections are linear and idempotent, , and a multivector is homogeneous of grade exactly when .
The sum in [eq:grade-sum] differs from the sum one learns first. Adding two vectors or two bivectors is addition within a single grade: the coefficients of the shared basis blades add and the result is again a vector or a bivector. Adding across grades is different. The sum of a scalar and a bivector does not collapse to one number or one plane. It stays a two-part object, as the complex number has a real part alongside an imaginary one that cannot be added into it. A multivector is therefore a list of independent grade slots, and the algebra is the direct sum
in which addition works coordinate-wise within each summand and never moves a term from one grade into another. Two multivectors are equal when they agree grade by grade. An identity between multivectors is therefore a family of identities, one per grade, read off by the projections of [eq:grade-sum]. So the geometric product's output of a scalar plus a bivector in [eq:geometric-decomp] is a single element of the graded sum: the scalar and the bivector occupy different slots of it.
This graded direct sum is the algebraic counterpart of the jet of Section 1.1. The multivector stacks the algebra's data by grade as the jet stacks a trajectory's data by order of differentiation, each a tower of independent levels. The grade projection reads one slot of the multivector as taking the -th derivative reads one slot of the jet.
The blades of all grades from an orthonormal basis, taken as ordered products of distinct basis vectors, are linearly independent and span everything the geometric product can build. Choosing of the basis vectors gives a grade- basis blade, and there are such choices. The whole algebra therefore has dimension
The alternating part of the algebra, the span of the outer products alone, is the exterior algebra of the same dimension . The geometric algebra contains it as a vector space and adds the metric information through the inner part of [eq:geometric-decomp].
The geometric product is fixed on vectors by the contraction axiom of [eq:contraction] and, being associative and bilinear, extends by distributivity to the whole graded algebra: any two multivectors are multiplied by expanding each into blades and multiplying the blades. This single associative product on multivectors is the geometric algebra, and the vector product of Section 1.6 is its grade- case. On homogeneous blades the product resolves by grade.
The geometric product of a grade- blade and a grade- blade spreads over the grades ; its lowest grade is the inner product and its highest grade is the outer product,
For two vectors, , these are the grade- scalar and the grade- bivector , so the decomposition of [eq:geometric-decomp] is recovered. The outer product raises grade and the inner product lowers it. The full geometric product does both at once and closes the algebra.
To make this precise without relying on orthonormal blades we give the defining property that characterises the algebra uniquely.
The geometric algebra is the associative unital real algebra, equipped with a linear inclusion , that is universal among algebras receiving with the contraction property: for every associative unital algebra and every linear map satisfying for all , there is a unique algebra homomorphism with .
The universal property fixes up to a unique isomorphism. Any construction that has it is the geometric algebra. One such construction proves that the object exists.
Let be the tensor algebra of , and let be the two-sided ideal generated by all elements . The quotient
is an associative unital algebra with the universal property of [clifford-universal], and it has dimension where , with the grade decomposition of [eq:dim-2n].
Proof
The tensor algebra is the free associative algebra on , so every linear extends uniquely to an algebra homomorphism . When satisfies the contraction property , that homomorphism annihilates every generator of the ideal and so factors through the quotient [eq:clifford-quotient] to a unique with , which is the universal property of [clifford-universal]. Imposing is exactly the contraction axiom of [eq:contraction]. Reducing every tensor to an ordered product of distinct orthonormal generators, using that relation on repeats and anticommutation on transpositions, leaves one basis blade for each subset of the generators, so the quotient is spanned by blades. These blades are independent because a -dimensional algebra with the contraction property exists: on the exterior algebra the operator squares to , and the induced homomorphism sends the blade to an operator taking to . No nontrivial combination of blades therefore vanishes. This gives the count and grade decomposition of [eq:dim-2n].
The geometric algebra is therefore the most general associative algebra in which vectors square to their lengths. We now fix the three-dimensional case the rest of the article inhabits.
1.8. Algebra
We read the general construction of Section 1.7 off at to give the algebra of physical space, in which the rest of the article computes. The label records the signature: three orthonormal directions each squaring to and none squaring to or . Its dimension is , split by grade as one scalar, three vectors, three bivectors and one trivector.
Fix an orthonormal basis of , so that and for . The eight basis blades of , by grade, are
where and the bivectors are written in the cyclic order that mirrors the cycle. The top-grade blade is the pseudoscalar, the unit oriented volume of space.
The whole algebra is fixed by how the basis blades multiply, and that in turn follows from and the anticommutation of distinct generators, worked out once and tabulated. The entry in row and column is the geometric product .
| · | 1 | e1 | e2 | e3 | e23 | e31 | e12 | I |
|---|---|---|---|---|---|---|---|---|
| 1 | 1 | e1 | e2 | e3 | e23 | e31 | e12 | I |
| e1 | e1 | 1 | e12 | −e31 | I | −e3 | e2 | e23 |
| e2 | e2 | −e12 | 1 | e23 | e3 | I | −e1 | e31 |
| e3 | e3 | e31 | −e23 | 1 | −e2 | e1 | I | e12 |
| e23 | e23 | I | −e3 | e2 | −1 | −e12 | e31 | −e1 |
| e31 | e31 | e3 | I | −e1 | e12 | −1 | −e23 | −e2 |
| e12 | e12 | −e2 | e1 | I | −e31 | e23 | −1 | −e3 |
| I | I | e23 | e31 | e12 | −e1 | −e2 | −e3 | −1 |
Three features of the table give the geometry. Each bivector squares to , as the diagonal entries show. A unit bivector therefore behaves like the imaginary unit, the algebraic root of the plane rotations that Section 1.5 found had no real eigenvalue. The pseudoscalar likewise squares to .
In the pseudoscalar satisfies
and it commutes with every element of the algebra. The centre of is the set and is a copy of the complex numbers.
Proof
The square is a short calculation. We run it up the dimensions to see where the value is fixed and where it changes sign. The one fact it uses is that distinct orthonormal vectors anticommute: their inner product vanishes, so the geometric product is the outer product alone, and the outer product is antisymmetric,
and the same holds for any with . Squaring a product of distinct basis vectors is then a matter of moving each repeated factor next to its twin, one sign per transposition, and collapsing .
The plane shows the mechanism with a single anticommutation. Its pseudoscalar is the unit bivector , and
Space has one factor more. Reversing the second copy of introduces the reversal sign: three transpositions give , after which the inner squares telescope from the middle outward,
which is [eq:I-squared]. The reversal works in every dimension the same way. For orthonormal vectors the pseudoscalar reverses with the sign of the transpositions that reverse the list, , and the telescoping of the inner squares leaves
The exponent is even when and odd when . Hence the pseudoscalar squares to in dimensions two and three alike, and again in six and seven, while it squares to in dimensions one, four, five and eight. That the plane and space share the value is why the single imaginary unit demanded by the plane rotation of Section 1.5 serves the rotations of both.
Commutativity is the remaining claim, and the same reversal count settles it. Moving a vector through the factors of yields . Hence the pseudoscalar is central exactly when is odd. Three-dimensional space qualifies, and the commutativity recorded in the last row and column of the table is this odd-dimensional fact at . Nothing else is central. The commutator of with a vector is the bivector and with a bivector the vector ; both vanish for every only when and . The centre is therefore .
That the centre is a copy of explains why the complex numbers keep appearing in three-dimensional rotation problems. They are already inside the real algebra of space, as the centre spanned by and the pseudoscalar . We compute one product in full to see the decomposition of [eq:geometric-decomp] at work.
Take the two vectors and and form by expanding bilinearly and reading each basis product from the table, one term at a time,
From the table , , and , so
The scalar part is and equals . The bivector part matches , whose components in the cyclic basis are with and . The single product of [eq:cl3-result] contains both the alignment and the oriented plane of and , exactly as [eq:geometric-decomp] requires.
Every later computation in the article is in principle a lookup in this table. The next section uses the pseudoscalar to explain and then retire the cross product.
1.9. Duality and the cross product
The bivector of Section 1.6 is the invariant object that the cross product only impersonated. This section makes the impersonation exact: multiplication by the pseudoscalar converts a bivector to the vector normal to it, and that operation is what the cross product is. The conversion needs the space of bivectors and the space of vectors to share a dimension. It works in three dimensions and nowhere else. This identification of the cross product with a vector is the dimensional accident.
The dual of a multivector in is its product with the inverse pseudoscalar,
the second equality follows because by [eq:I-squared]. Duality exchanges grades and : it sends vectors to bivectors and bivectors to vectors, scalars to pseudoscalars and back. It is the geometric-algebra form of the Hodge star, and since is central the side of multiplication does not matter.
Duality takes the three basis bivectors of [eq:cl3-basis] to the three basis vectors. The last column of the multiplication table gives the correspondence in the cyclic labelling: , and . The oriented plane , for instance, is dual to the vector normal to it with the right-handed sense. This is the exchange that Section 1.5 flagged as buried inside the cross product, now performed in the open.
In the cross product of two vectors is the dual of their outer product,
The identity holds in three dimensions alone, because it needs the space of bivectors and the space of vectors to have equal dimension so that can move between them. In the plane there are no vectors dual to the single bivector and in four dimensions the dual of a bivector is another bivector. The pseudovector character of noted in Section 1.5 is now explained: itself changes sign under reflection and the dual of a bivector therefore reflects with the extra sign that makes it a pseudovector.
Proof
Expand both sides in the cyclic basis. The outer product is , and dualising each basis bivector by , and sends it to , which are the components of . Since is central, . The step turns three basis bivectors into three basis vectors, and this pairing of equal counts occurs at alone.
A bilinear vector-valued product with the length property exists in three dimensions and also in seven, where it is inherited from the multiplication of the octonions. What is special to three dimensions is the identification of [eq:cross-dual], the reading of that product as the dual of a bivector. In seven dimensions the outer product is a bivector in a twenty-one-dimensional space whose dual is a five-blade. The seven-dimensional cross product therefore exists but is not the dual of a plane and lies outside the correspondence developed here. The dimensional accident is the three-dimensional identification.
The moving figure shows the bivector and its dual normal together. As the plane tips the vector follows it in three dimensions only.
Wherever mechanics traditionally writes a cross product, as for an angular velocity, a torque or an angular momentum, we shall write the bivector , of which [eq:cross-dual] shows the cross product to be a disguise. The bivector is defined in every dimension, reflects faithfully and needs no handedness convention. The pseudovector complaints of Section 1.5 therefore lapse. The dual remains available whenever a normal vector is wanted but is no longer the primary object. Next we treat rotations and the elements of the algebra that generate them.
1.10. Reflections and rotors
Rotations are the motions of rigid-body mechanics and Section 1.5 found the matrix picture of them wanting: complex eigenvalues in the plane and an axis that is an accident of dimension. Geometric algebra generates rotations from a more primitive motion, the reflection, and represents them by single algebra elements, the rotors, which compose by multiplication and never lock. This section builds both, works a rotor in full and records the double cover of the rotation group they realise.
A reflection is expressed by a sandwich of a vector between a unit vector and its inverse.
Let be a unit vector, , so that . The map
is the reflection of in the plane through the origin with normal , reversing the component of along and fixing the plane. In components it equals .
Proof
Split into parts along and across . The parallel part commutes with and the perpendicular part anticommutes with it, by the two special cases of Section 1.6, so , which reverses the normal component and fixes the plane. Writing the parallel part as gives the component form .
Now compose two reflections. Reflecting in the plane normal to and then in the plane normal to gives
and the product of two unit vectors is a new kind of element that effects a rotation. Two reflections make a rotation, and the angle of that rotation is twice the angle between the mirrors.
A rotor is a product of an even number of unit vectors, equivalently an even-grade element with scalar and a bivector, normalised so that , where the reverse reverses the order of all vector factors. A rotor acts on a vector by the sandwich
which is a rotation. For a unit bivector with , the rotor of a rotation through angle in the plane of is the exponential
the series collapsing to cosine and sine because . Products of rotors are rotors. In the general theory a product of any number of unit vectors is called a versor, which generates an orthogonal transformation by the sandwich; rotors are exactly the even versors and reflections the odd ones. Here the rotors form the group we need.
The exponential of [eq:rotor-exp] makes the bivector the direct generator of the rotation in place of the antisymmetric matrix or the axis vector of the matrix language. The plane of turning is itself, present in every dimension, and the half-angle comes from the two-reflection construction of [eq:two-reflections]. The figure spins a frame under the sandwich and sets the composition of two rotors beside the gimbal-lock path of Euler angles.
Rotate the vector through in the plane, whose unit bivector is with . The rotor of [eq:rotor-exp] is
Apply the sandwich [eq:rotor-sandwich] in two steps. First , using from the table. Then
From the table and , so the bracket is . Hence . The vector has turned into , a quarter turn in the plane, exactly as a rotation through should. The same computation with a general angle sends to .
The rotor and its negative give the same rotation, since . Running the angle from to takes from to while the frame returns to where it started. The rotors of unit norm therefore form a group that wraps twice around the rotation group .[2]
The map sending a unit rotor to the rotation is a two-to-one group homomorphism from , the group of unit even-grade elements of , onto the rotation group , with and mapping to the same rotation. As a manifold is the unit sphere in the four-dimensional even subalgebra, so . It is the reason a spin-half object must turn through to return to itself.
Proof
The sandwich preserves the inner product, since gives , and it keeps vectors in grade one, so it is an element of ; being a product of an even number of reflections it has determinant and lies in . The map is a homomorphism because composes the two sandwiches, it is onto because every rotation of space has an axis and is the product of two reflections in planes through that axis meeting at half the rotation angle, which [eq:two-reflections] writes as a rotor sandwich, and its kernel is the unit scalars , so and have one image and the map is two-to-one. The unit even elements with fill the unit sphere of the four-dimensional even subalgebra, which is .
The reflection figure composes two mirror operations into a single rotation, the geometric content of [eq:two-reflections].
Part III integrates rotors to follow a spinning body free of Euler angles and gimbal lock. One section of algebra remains: how a linear operator of Section 1.4 acts on the higher-grade blades, where the determinant finally becomes a product.
1.11. Operators in geometric algebra
A linear operator of Section 1.4 acts on vectors. Geometric algebra has bivectors, trivectors and the pseudoscalar as well, and there is one canonical way to extend an operator to all of them at once. The extension is the outermorphism. It settles the last discomfort of Section 1.5: the determinant, bolted on from outside in the matrix language, becomes the action of the operator on the pseudoscalar and is a product within the algebra.
Let be a linear operator. Its outermorphism is the unique extension of to all of that fixes scalars, agrees with on vectors and preserves the outer product,
It is grade-preserving, sending each -blade to a -blade, and it respects composition, .
The outermorphism is the natural way an operator drags oriented areas and volumes along with the vectors, since a blade is an oriented volume element and [eq:outermorphism] transports each of its spanning vectors. Applied to the top blade, the pseudoscalar of [eq:cl3-basis], it can only rescale it because the trivectors of are one-dimensional. The scale factor is the determinant.
For a linear operator on the outermorphism acts on the pseudoscalar by
and this equation may be taken as the definition of . It is manifestly basis-independent, it gives the multiplicative law at once from , and its sign reports whether preserves or reverses the orientation defined by .
Proof
The trivectors of form a one-dimensional space, so the grade-preserving can only rescale the top blade, and is a scalar multiple of ; [eq:det-pseudoscalar] names that scalar . Expanding by [eq:outermorphism] and collecting the antisymmetric coefficients reproduces the permutation sum , the classical determinant, so the two definitions agree. Basis-independence is immediate because neither nor [eq:det-pseudoscalar] names a basis, and gives .
The determinant is now a statement about volume inside the algebra: is the factor by which inflates the unit oriented volume . The figure morphs a grid and a volume cell under an operator and reads the determinant off as the volume scale.
The inner product supplies one further operator, the adjoint, which the inertia tensor of Part III will need.
The adjoint of a linear operator is defined through the inner product of Section 1.3 by
In an orthonormal basis its matrix is the transpose, . An operator is symmetric when and orthogonal, a rotation or reflection, when , so that it preserves every inner product.
Where possible an operator is diagonalised by finding the blades it only rescales.
A nonzero blade of grade is an eigenblade of with eigenvalue when . A grade-one eigenblade is an ordinary eigenvector. A symmetric operator on has three mutually orthogonal eigenvectors with real eigenvalues, and the product of the three eigenvalues is by [eq:det-pseudoscalar], since scales the volume spanned by the eigenvectors by each eigenvalue in turn.
This closes the algebra of a single instant. The Euclidean space of Sections 1.1 through 1.4 has grown through the geometric product into the algebra with its bivectors, rotors, duality and outermorphisms, and every discomfort of Section 1.5 has been met. It remains to let the instant move and to differentiate the vector-valued functions of time and space of which mechanics is made. Part II does this.
Part II
2. Two calculi
2.1. Vector-valued functions and the Fréchet derivative
Mechanics is written in vector-valued functions. A trajectory is a map from an interval of time to , a force field a map from to , a change of coordinates a map from one copy of to another. To differentiate any of them we need a derivative for a map between vector spaces. This is the best linear approximation to the map near a point. It is a linear map of the kind Section 1.4 gave a matrix. The derivative is the linear map that the partials represent.
A map is differentiable at a point when there is a linear map , its derivative, for which the increment splits into a linear part and a remainder that dies faster than the displacement,
The linear map is unique when it exists and does not depend on any choice of basis.
By Section 1.4 the linear map of [eq:frechet] has a matrix once bases are fixed. Read off through the entries of [eq:matrix-entries], this matrix is the Jacobian.
In the standard bases of and , the matrix of the derivative of [eq:frechet] is the Jacobian,
whose -th column is the partial-derivative vector at . When the Jacobian is square and its determinant is defined.
The derivative is one object. Its Jacobian [eq:jacobian] is its basis-dependent shadow, as the column [eq:column-vector] is the shadow of a vector and the matrix [eq:matrix-entries] the shadow of an operator. Linearity gives the chain rule: the derivative of a composite is the composite of the derivatives, so its Jacobian is the matrix product , the composition law of Section 1.4 transcribed for derivatives. We compute one Jacobian and read its determinant as a volume scale, as for the outermorphism of Section 1.11.
The polar map sends polar coordinates in the plane to Cartesian ones. Its partials fill the columns of [eq:jacobian]: differentiating each component with respect to and then ,
Its determinant is
which is the factor by which the map inflates an area element, the outermorphism scale [eq:det-pseudoscalar] of Section 1.11 in dimension two. It is the that weights the polar area element. It vanishes at the origin, where the map folds the whole -line onto one point and ceases to be invertible. The determinant of the derivative is the local measure that Section 2.4 integrates.
2.2. Vector derivative
The Jacobian of Section 2.1 collects every partial derivative into an array. Geometric calculus assembles the same partials into a single object of the algebra, a vector whose components are the derivative operators . As a vector it multiplies fields by the geometric product of Section 1.6, and that product unifies the differential operators of vector analysis in Section 2.3.
On with a basis and its reciprocal basis fixed by , the vector derivative is the grade-one operator
It acts on a multivector field by the geometric product, one basis direction at a time,
In an orthonormal basis the reciprocal basis coincides with the basis, , and .
The reciprocal basis makes [eq:vector-derivative] coordinate-free: under a change of basis the vectors and the covectors transform inversely, so the sum is the same operator in every frame. It is a vector in the sense Section 1.2 drew between an object and its representation. Contracting with a fixed direction gives the rate of change along it,
the directional derivative, the derivative of Section 2.1 evaluated on the vector , since .
The grade of is one. Its geometric product with a field therefore has the two-part structure of [eq:geometric-decomp]: an inner part that lowers grade and an outer part that raises it. On a scalar field there is no lower grade to reach, so the product is the grade-raising part alone, a vector, the gradient
which points along the steepest ascent of and whose magnitude is that slope. On a vector field both grades exist, and the two parts are the divergence and the curl of the next section.
2.3. Vector-analysis operators unified
The geometric product of two vectors splits as by [eq:geometric-decomp] into a scalar and a bivector. With the vector operator of Section 2.2 as first factor the scalar part is the divergence and the bivector part the curl. Together they are the whole first-order derivative of a vector field.
For a vector field on the geometric product decomposes into its scalar and bivector parts,
the scalar part being the divergence and the bivector part the curl,
With the gradient [eq:grad-def] on scalar fields the three operators of vector analysis are grade parts of the single product .
Proof
Since is a grade-one operator its product with a vector field splits term by term as in [eq:geometric-decomp]. The scalar part is , the divergence, and the bivector part is , the curl. These are the two sums of [eq:div-curl]. The split uses only that the first factor is a vector.
The curl of [eq:div-curl] is a bivector, the oriented-plane object of Section 1.6, and is defined in every dimension. The familiar curl vector of three-dimensional vector analysis is its dual: by the duality of Section 1.9 and the relation [eq:cross-dual] one has . The classical curl is the bivector seen through the pseudoscalar and inherits the pseudovector character flagged in Section 1.5. The bivector is primary and the axial vector is its three-dimensional shadow. We compute the whole product for a field.
Take the vector field and form in the orthonormal basis. Differentiating the field along each axis, one direction at a time,
and multiplying each on the left by its basis vector,
Reading each basis product from the table of Section 1.8, with , , , and , the six terms collect into a scalar and a bivector,
The scalar part is the divergence , and the bivector part is the curl . Dualising the curl by [eq:cross-dual], with , , and , recovers the classical curl vector . To check, the component predicted by [eq:div-curl] is , matching the coefficient in [eq:nabla-F-result], and the divergence matches likewise. The single product [eq:nabla-F-result] contains the divergence and the whole curl.
2.4. Fundamental theorem of geometric calculus
One operator has one integral theorem. It states that integrating the vector derivative of a field over a region equals integrating the field itself over the boundary, once each integral is taken against the oriented measure the algebra supplies. The named theorems of vector calculus are its grade parts: Green's theorem, the divergence theorem and Stokes' theorem. The ordinary fundamental theorem of calculus is its one-dimensional case.[1]
Let be a compact oriented -dimensional region with boundary . Let be the directed measure on , the oriented unit -blade tangent to scaled by the ordinary volume element, and the directed measure on , the oriented -blade of its boundary. Then for any multivector field ,
where differentiates the field alone and, when , stands for the vector derivative of itself, the projection of the ambient onto the tangent space of . When is a region of of full dimension , the tangent blade is the constant pseudoscalar of Section 1.8, so .
Proof sketch
Subdivide into small oriented cells. On each cell the directed integral of is to first order the difference of across opposite faces, since is the first-order change of read against the directed measure. Summing over the cells cancels every interior face, because two adjacent cells share it with opposite boundary orientation, and leaves only the outer faces, whose sum is . Refining the subdivision sends the first-order error to zero and gives [eq:ftgc]. The one-dimensional case is the fundamental theorem of calculus and the general statement is the multivector form of Stokes' theorem.[1]
[eq:ftgc] states that the vector derivative and the boundary operation are adjoint under integration: differentiating inside is the same as restricting to the edge. Every classical integral theorem is [eq:ftgc] read at a particular grade. Take a solid region of and a vector field, so that and the boundary measure is with the outward unit normal. Multiplying [eq:ftgc] on the left by and keeping the scalar part turns into its divergence [eq:div-curl] and into . The result is the divergence theorem
On a surface with a vector field the scalar part pairs the oriented area element with the bivector curl of [eq:div-curl] and the boundary element with . That scalar identity is Stokes' theorem. The same construction in the plane gives Green's theorem. On an interval with a scalar field [eq:ftgc] is the fundamental theorem of calculus . One theorem, projected onto each grade in each dimension, is the whole zoo. We test it in the simplest planar case.
Work in the plane with the radial field and let be the unit disk. Its divergence is , a constant, so the left side of [eq:divergence-theorem] is twice the area of the disk. On the boundary circle the outward unit normal is the position vector itself, , so the flux integrand is on , and the right side is the circumference. Both sides agree,
The area integral of the divergence equals the boundary flux, as [eq:divergence-theorem] and, above it, [eq:ftgc] require.
2.5. Matrix-calculus zoo versus one theorem
This section is the second thesis anchor of the article. Section 1.5 catalogued the discomforts of the matrix language for the algebra of a single instant. Here we list the defects of the classical apparatus for the calculus and set against them the single operator and single theorem of geometric calculus.
The classical vector calculus meets each geometric task with a bespoke tool. It has three first-order differential operators with incompatible type signatures: the gradient sends a scalar field to a vector field, the divergence sends a vector field to a scalar field, and the curl sends a vector field to another vector field. The three cannot be composed freely, since the output of one is rarely a legal input to the next. Each is proved and remembered on its own. The curl vector exists in three dimensions alone. It is built on the same trade of an oriented plane for its normal that made the axial cross product a dimensional accident in Section 1.9. In the plane there is no curl vector and in four dimensions no vector to represent it, though the bivector curl is defined in every dimension. The integral calculus is as fragmented. The fundamental theorem of calculus governs an interval, Green's theorem a planar region, the divergence theorem a solid, and Stokes' theorem a surface, four statements with four proofs, each welded to its dimension and its grade.
Geometric calculus replaces the whole apparatus with one operator and one theorem. The vector derivative of [eq:vector-derivative] is a single grade-one object. By the geometric product it gives the gradient on a scalar field and on a vector field the divergence and curl together as the scalar and bivector parts of [eq:nabla-decomp]. The fundamental theorem [eq:ftgc] is a single statement with Green's, the divergence and Stokes' theorems as grade projections, as Section 2.4 showed. The table lists the correspondence.
| Classical apparatus | Type and range | Geometric calculus |
|---|---|---|
| gradient | scalar to vector, any | on a grade- field |
| divergence | vector to scalar, any | scalar part |
| curl | vector to pseudovector, only | bivector part , any |
| calculus, Green, Gauss, Stokes | four theorems, four proofs | one theorem [eq:ftgc] |
The gain matches the first anchor. There the geometric product absorbed the inner product, the cross product and the determinant into one associative multiplication. Here the vector derivative absorbs the gradient, divergence and curl into one operator and the fundamental theorem absorbs four integral theorems into one identity. In both cases the apparatus that the matrix language spreads across special cases and special dimensions consists of grade parts of a single object defined in every dimension and needing no handedness convention. The figure shows the vector derivative on a flowing field, with its scalar divergence and bivector curl as the field deforms and the boundary flux of [eq:ftgc] accumulating in step.
The two calculi are now level, as the two languages were at the close of Part I. Position, its derivative and the operator that differentiates fields are in hand in both tongues. Part III sets the machinery in motion: the position vector of Section 1.1 acquires a mass and a law, and the bivector of Section 1.6 becomes the angular momentum of a spinning body.
Part III
3. Newtonian and rigid-body dynamics
3.1. Kinematics
Physics begins where the algebra and calculus of the first two parts are set in motion. Section 1.1 gave the trajectory, velocity and acceleration: the position vector depends on time and its first two derivatives are the arrows tangent to the path. Here we take them up in the coordinate space of the dynamics, recall the one estimate they rest on and pass to the momentum that Newton's laws govern.
As in Section 1.1 a trajectory is a smooth vector-valued function of time, and its velocity and acceleration are its first and second time derivatives,
each a vector in the same space as the position, with the overdot abbreviating throughout.
The one-parameter derivative of [eq:velocity-acceleration] is the limit that the increment bound of [increment-bound] controls. It is the one-parameter case of the Fréchet derivative of [eq:frechet]: with a single scalar input the linear map is multiplication by the vector , so the derivative and the velocity are one object. Since differentiation acts componentwise on a vector-valued function the velocity in an orthonormal basis is , the basis vectors being constant in an inertial frame. That clause is all that inertial means for kinematics. Section 3.5 studies what the velocity becomes when the basis turns.
Two derived quantities describe the motion. The speed is the length of the velocity, built from the inner product of Section 1.3. Differentiating gives , so the acceleration is orthogonal to the velocity exactly when the speed is momentarily stationary. The bivector of Section 1.6 is the oriented plane of the motion, and it vanishes precisely on straight-line paths, where velocity and acceleration are parallel. These two invariants, the scalar and the bivector , are the grade parts of the single geometric product and give the curvature and turning of the path without a cross product.
3.2. Newton's laws
Kinematics describes how a body moves. Dynamics says what makes it move so. The bridge is the mass, a positive scalar attached to the body, and the law that ties the rate of change of its motion to the force upon it. This section states Newton's three laws in the vector language of Section 1.1, with the momentum as the primary dynamical variable, since the second law governs and the third law conserves the momentum.
The mass of a body is the scalar measuring its resistance to being accelerated. Its momentum is the vector
the mass times the velocity of [eq:velocity-acceleration], pointing along the direction of motion with magnitude .
The inertia of a body is its resistance to any change in its momentum, and the scalar that measures that resistance is the mass of [eq:momentum]. One and the same force changes the momentum of a heavy body and a light one alike but alters the velocity of the heavy one the less. Newton's first law is the law of inertia: a body under no force keeps its momentum fixed. The inertial frames of Section 1.1 are those in which it holds.
The word records this idea in its history. Momentum is Latin for moving power, a contraction of movimentum from movere, to move, and shares a root with moment. The kinship is exact. A moment in mechanics is the oriented plane that a quantity spans with the position vector, its wedge with . The torque of the next section is the moment of the force and the angular momentum is the moment of the momentum, each the plane the position sweeps with the vector it turns. The linear momentum of this section and the angular momentum of the next are one construction read at two grades. The variational chapters name the same quantity a third time as the conjugate momentum of [conjugate-momentum], the measure of motion belonging to a generalised coordinate. Each guise has its own conservation condition: the linear momentum when no force acts, the angular momentum when no torque acts, and the conjugate momentum when its coordinate is absent from the Lagrangian.
Newton's three laws are statements about momentum in an inertial frame. The first law asserts that such frames exist: there is a class of frames, the inertial ones of Section 1.1, in which a body subject to no force keeps its momentum constant and traces a straight line at constant speed. The first law is the precondition of the second and fixes the arena in which the second law is read.
In an inertial frame the force on a body is the vector equal to the time rate of change of its momentum,
which for constant mass reduces to with the acceleration of [eq:velocity-acceleration]. Force and acceleration are proper vectors of , adding by the parallelogram rule of Section 1.2.
The equation [eq:second-law] is a vector equation, three scalar equations at once, and inherits from Section 1.2 the fact that forces superpose: the total force is the vector sum of the individual forces and the single resultant drives the momentum. The third law completes the account for interacting bodies: the force exerted by a first body on a second is met by an equal and opposite force of the second on the first, . Hence the total momentum of an isolated pair or of any isolated system is conserved. Summing [eq:second-law] over the parts of such a system cancels every internal force in pairs and leaves , the momentum form of the law that governs the whole. The same construction one grade higher gives the angular momentum of the next section.
3.3. Angular momentum as a bivector
The rotational analogue of momentum is where the matrix language first shows the defects catalogued in Section 1.5. Tradition writes the angular momentum as the cross product , a pseudovector that flips with the wrong sign under reflection and can be formed as a vector only in three dimensions. The invariant object is the one the cross product was impersonating: the oriented plane swept by the position and the momentum, the bivector of Section 1.6. This section takes the bivector as the definition. Its law is cleaner than the pseudovector's.
The angular momentum of a body about the origin is the bivector
the oriented plane spanned by the position and the momentum, with magnitude the area of the parallelogram they span. The torque of a force about the origin is likewise the bivector
the oriented plane of the position and the force.
Both quantities are grade-two elements of , defined by the outer product alone, so they have no hidden handedness and reflect as faithfully as the plane they name. The pseudovectors and of the matrix account are their duals, and so on through [eq:cross-dual], the normal vectors read off the planes by the pseudoscalar of Section 1.9. The plane is primary and the axial vector its three-dimensional shadow. The law relating the two is the rotational counterpart of the second law [eq:second-law] and follows from it in one line.
For a body moving under a total force by the second law [eq:second-law], the angular momentum bivector [eq:angular-momentum] and the torque bivector [eq:torque] satisfy
Proof
The proof differentiates the outer product and uses the antisymmetry of the wedge. Differentiating [eq:angular-momentum] by the product rule, which the bilinear outer product obeys,
The first term vanishes: , since the outer product of any vector with itself is zero by the antisymmetry of Section 1.6. The second term is by the second law, the torque [eq:torque]. So , exactly [eq:dL-dt-tau]. The vanishing of the velocity term is the geometric fact that a body cannot exert torque on itself by moving along its own line of motion. The wedge records this automatically where cross-product bookkeeping invokes the parallel-vector rule by hand.
The figure shows the oriented plate of alongside the orbiting body.
The corollary is the conservation law of the next section. When the torque [eq:torque] vanishes the angular momentum bivector [eq:angular-momentum] is constant in time, plane and magnitude both fixed. A vanishing torque means the force lies along the position vector. Forces of that kind are the central forces.
3.4. Central forces and Kepler's areal law
A central force is one directed always along the line joining the body to a fixed centre, its magnitude depending only on the distance. Gravity and the Coulomb force are of this kind, and the two-body problem lives here. The bivector law [eq:dL-dt-tau] of the previous section makes the central case transparent: the torque vanishes identically, the angular momentum plane is frozen, and Kepler's second law drops out as the statement that the frozen bivector is swept at a constant rate. This section derives that areal law and checks it on a worked orbit.
A central force about the origin is a force of the form
directed along the unit position vector , with scalar magnitude a function of the distance alone. It is attractive where and repulsive where .
Because the force [eq:central-force] is parallel to the position, its torque bivector [eq:torque] vanishes, and the rotational law [eq:dL-dt-tau] then freezes the angular momentum.
Under any central force [eq:central-force] the angular momentum bivector [eq:angular-momentum] is constant,
In particular the motion is confined for all time to the fixed plane of .
Proof
The torque is , the outer product of a vector with a multiple of itself. Hence the rotational law [eq:dL-dt-tau] gives and the constancy [eq:L-const]. For the confinement, because the outer product repeats the factor , and a vector whose wedge with a fixed nonzero bivector vanishes lies in that bivector's plane. Hence the orbit stays in the plane of . When the velocity stays parallel to the position and the motion runs along a fixed line through the centre. Any plane containing that line contains the orbit.
The pseudovector account states the planarity of the orbit as the constancy of an axis. Here it is the constancy of the plane itself. The conserved magnitude gives Kepler's law. Write the bivector as . The quantity is the oriented area swept per unit time by the radius vector, since in a time the radius sweeps the triangle of oriented area .
The areal velocity of a trajectory is the oriented area swept per unit time by the position vector,
a bivector equal to the angular momentum divided by twice the mass.
Kepler's second law is now immediate: is constant by [eq:L-const], so the areal velocity [eq:areal-velocity] is constant and the radius vector sweeps equal oriented areas in equal times. The law that Kepler inferred from Tycho Brahe's tables of Mars states that the bivector does not change, plane or magnitude. It holds for every central force. We test it on a circular orbit, where every quantity is elementary.
Let a body of mass move on the circle of radius at constant angular rate in the plane,
which is the path a central force supplies when its inward pull matches the required acceleration. Differentiating, the velocity is . Form the areal bivector one product at a time,
Expanding the wedge and using and , the cross terms combine through to give
a bivector independent of , so the areal velocity [eq:areal-velocity] is the constant and the angular momentum is . The time to sweep the whole disc is the period , and the area swept is , the area of the circle, as it must be. The constancy of [eq:circle-result] is Kepler's second law in the one case where it can be read by inspection. In a general orbit and the angular rate vary while their combination in [eq:areal-velocity] stays fixed.
3.5. Rotating frames and the Coriolis bivector
The inertial frame of Section 1.1 was chosen so that the basis vectors stay fixed and the velocity of Section 3.1 is a plain componentwise derivative. Much of mechanics is done in a frame that turns, such as the rotating Earth. There the basis vectors are functions of time. The rotor of Section 1.10 takes one frame to another. Letting it depend on time produces the angular velocity as a bivector and the Coriolis and centrifugal terms as consequences.
Let a time-dependent rotor relate the fixed inertial basis to the rotating basis through the sandwich of [eq:rotor-sandwich]. Since at every instant differentiating that constraint pins the rate of the rotor to a bivector.
For a time-dependent rotor , the angular-velocity bivector in the space frame is
Differentiating gives , so equals minus its own reverse and is therefore a pure grade-two element, a bivector, with no scalar part. Its body-frame form is the transported bivector .
That is a bivector is the rotational counterpart of the fact that velocity is orthogonal to position on a sphere: the rotor lives on the unit sphere of the even subalgebra of Section 1.10 and its velocity is tangent there. This is the grade-two condition. The single bivector replaces the antisymmetric matrix and the axial angular-velocity vector of the matrix account at once and names the plane of turning directly, in the manner of [eq:rotor-exp].
The bivector governs how the time derivative seen in the rotating frame differs from the one seen in the fixed frame. Let be the body-frame description of a body's position, so that its space position is . Differentiating that sandwich gives the transport rule for the space velocity seen in the rotating frame.
For any body-frame vector the space-frame velocity expressed in the rotating frame exceeds the naive rotating-frame rate by a term built from the angular-velocity bivector,
where is the grade-one contraction of the vector with the bivector, the geometric-algebra form of the term written in the matrix language.
Proof
Differentiate by the product rule, . Refer the result to the rotating frame by the sandwich and use from [eq:angular-velocity], so that and . The first and third terms combine into , the grade-one contraction of the vector with the bivector, and the middle term is . Together they give [eq:transport-rule].
Applying the rule [eq:transport-rule] a second time to the velocity unfolds the acceleration into four terms, since each of the two summands in [eq:transport-rule] is itself transported and the bivector may vary. Newton's second law [eq:second-law] written in the rotating frame then has three extra terms beside the true force,
with the true force referred to the rotating frame. The three added terms are the inertial forces of the turning frame, each a contraction against the angular-velocity bivector. The Coriolis term acts on a body only while it moves within the frame and turns its path sideways in the plane of ; the centrifugal term points outward from the axis of the plane and grows with distance from it; and the last term, present only when the rotation rate itself changes, is the azimuthal or Euler force. The figure deflects a body crossing a spinning platform and marks the plane of in which the deflection lies.
The Coriolis term deflects the trade winds and swings the plane of a pendulum, and the bivector shows its geometry: the deflection lies in the plane of rotation and is proportional to the component of the velocity lying in that plane. No axis and no handedness convention enter. In the next section the rotor that generated as an external frame becomes the orientation of a rigid body.
3.6. Rigid-body rotor kinematics
This section is the third thesis anchor of the article. A rigid body is a collection of points whose mutual distances never change, so its configuration is fixed by the position of one reference point and the orientation of the body. The orientation is the difficulty. The matrix language breaks down there and the language of Euler angles worse. We represent the orientation by a rotor, whose equation of motion is a single first-order bivector law that never locks and composes by multiplication.[3]
Fix a reference configuration of the body and label each of its points by the constant vector it occupies there. At time the body has turned, and the same material point sits at the space position obtained by the rotor sandwich of [eq:rotor-sandwich],
so that a single rotor , the same for every point, gives the entire orientation of the body. The rigidity is automatic: the sandwich is a rotation and rotations preserve every inner product by Section 1.11, so all mutual distances and angles are fixed. The motion of the whole body reduces to the motion of one rotor on the group of Section 1.10.
The orientation rotor of a rigid body evolves by the first-order bivector law
where is the space-frame angular-velocity bivector of [eq:angular-velocity]. Equivalently in terms of the transported body-frame bivector , and either form preserves the normalisation exactly.
The law [eq:rotor-equation] is the relation [eq:angular-velocity] that defined the angular velocity, read as an evolution equation for the rotor once is supplied by the dynamics of the following sections. Its exact flow keeps the orientation a proper rotation for all time.
Proof
That the law preserves the normalisation is a one-line check. Differentiate and use [eq:rotor-equation] together with the fact that the reverse of a bivector is its negative, ,
The exact flow therefore stays on , as the matrix equation stays on the orthogonal group. The two differ under a numerical step, which leaves either object slightly off its group. A drifted rotor is a nonzero even element whose sandwich is still a rotation scaled by , and one division by restores it exactly. A drifted matrix has lost orthogonality in all nine entries and needs a Gram-Schmidt or polar reorthogonalisation.
The matrix language describes orientation by three Euler angles. The equations relating their rates to the angular velocity contain a division by the sine of the second angle, which vanishes when two of the rotation axes align. There the map from angles to orientation degenerates, two of the three freedoms collapse into one and the equations of motion blow up though the body does nothing singular. This is gimbal lock, an artefact of coordinatising the orientation by angles. The rotor law [eq:rotor-equation] has no such defect. It is a single equation on the whole group with no coordinates on the orientation and so no coordinate singularity, and its right side is a smooth bivector at every configuration. Composing two orientations is the product of their rotors, associative and closed by Section 1.10. Euler angles compose by an awkward non-commutative table and rotation matrices have nine numbers under six constraints. One rotor has four numbers under one constraint and obeys one bivector equation that never locks.
The rotor law also completes for rotation the jet construction that Section 1.1 built for translation. A trajectory has its position, velocity and acceleration in the two-jet of [two-jet]. An orientation has the corresponding data in the successive derivatives of its rotor.
The rotational -jet of a motion is the tuple of the orientation rotor and its first time derivatives,
each entry an element of the even subalgebra of Section 1.10. Its first level is the angular velocity: by [eq:angular-velocity] the pair fixes the bivector , exactly as the pair fixed the velocity. Its second level fixes the angular acceleration in the same way.
The angular acceleration is the time rate of the angular-velocity bivector, and the second rotational jet [eq:rotational-jet] delivers it as the grade-two part
a bivector at every instant, the rotational counterpart of the acceleration that the second entry of the translational jet supplies.
Proof
Differentiate the angular velocity of [eq:angular-velocity] by the product rule,
The reverse commutes with differentiation, , and the rotor law of [eq:rotor-equation] gives through . Hence . A bivector squares to a scalar, so is grade zero and the entire term is scalar. Now is the derivative of a bivector-valued function and so is pure grade two. That scalar therefore cancels the scalar part of and leaves , as in [eq:angular-acceleration].
The two jets now run in parallel, the translational and the rotational , and the scale-graded norm of [scale-graded-norm] weighs the levels of either against a displacement of size alike. The estimates that control a trajectory therefore control an orientation with no new analysis. This is the parity between the jet and the graded algebra drawn there for multivectors, applied to the group of rotations.
The dynamics needs the relation between the angular velocity that drives [eq:rotor-equation] and the angular momentum that the torque changes. That relation is the inertia operator on bivectors.
3.7. Inertia operator on bivectors
The angular velocity of Section 3.6 is a bivector and by Section 3.3 so is the angular momentum. The inertia of the body relates them, and its natural home is a linear operator on the grade-two space of bivectors, of the kind Section 1.11 extended to blades. This section builds that operator and derives from it the angular momentum and the kinetic energy.
The classical form of the inertia is the moment of inertia about an axis, a single number that the operator generalises. It is a second moment of the mass in the sense of moment given in Section 3.2: the angular momentum is the first moment of the momentum and the moment of inertia is the second moment of the mass about the axis of turning.
The moment of inertia of a body about an axis through the origin, with unit direction , is the mass-weighted sum of squared perpendicular distances to the axis,
with the distance of the mass element from the axis. For a continuous body the sum is the integral .
When a body turns about a fixed axis with angular-velocity bivector , where is the unit plane orthogonal to the axis, its angular momentum in that plane and its kinetic energy are
with the moment of inertia [eq:moment-of-inertia] about the axis.
Proof
A point at body position turning with has velocity , whose length is , since the contraction with the unit plane returns the in-plane component of , of length , turned a quarter turn. Its kinetic energy is , and summing over the body gives by [eq:moment-of-inertia]. The angular momentum has in the plane each term equal to , so . Both are the special case of the operator relations built next for a principal axis, where the plane is an eigen-bivector with eigenvalue .
A freely turning body presents a different moment of inertia to every plane at once. One linear map packages them all. Consider a rigid body turning about a fixed point, so that each material point at body position has velocity by the contraction of Section 3.5. Summing the angular momentum [eq:angular-momentum] over the body, with mass elements at positions , gives . This sum is linear in the bivector . That linear dependence is the inertia operator.
The inertia operator of a rigid body about a fixed point is the linear map on bivectors
sending a grade-two element to a grade-two element, with the sum running over the mass elements at body positions . For a continuous body the sum is the mass integral .
The operator [eq:inertia-operator] is the geometric-algebra form of the inertia tensor, a symmetric operator in the sense of the adjoint of [eq:adjoint], now expressed with the bivector scalar product that pairs two bivectors to a scalar. Its symmetry, , follows from the manifest symmetry of [eq:inertia-operator] under exchange of and in the sum, and by the eigenblade proposition of Section 1.11 guarantees three mutually orthogonal eigen-bivectors with real eigenvalues. The eigenvalues are positive unless the whole mass lies on one line through the fixed point.
Being symmetric the inertia operator [eq:inertia-operator] has three orthogonal principal eigen-bivectors with positive eigenvalues, the principal moments of inertia ,
each eigen-bivector being the plane orthogonal to a principal axis of the body. In the body frame aligned with these planes the operator is diagonal, and a general angular velocity has angular momentum .
The two dynamical quantities of the body follow from the operator. The angular momentum is its value on the angular-velocity bivector, and the kinetic energy is half its associated quadratic form.
For a rigid body turning with angular-velocity bivector the angular momentum bivector and the kinetic energy are
the energy a positive-definite quadratic form in the components of along the principal planes, since each moment .
Proof
The first equation of [eq:L-inertia] is the summed angular momentum computed above. The second is the summed kinetic energy , which the symmetry of the operator collects into the single scalar product . Both are the bivector statements of quantities the matrix language writes with the inertia matrix acting on an axial angular-velocity vector. Here the operator acts within the grade-two space throughout.
The figure integrates a free top under the rotor law [eq:rotor-equation] and displays the angular-velocity and angular-momentum bivectors as oriented plates. They differ unless is a principal eigen-bivector.
That and point along different planes unless the body turns about a principal axis is the source of the behaviour of a spinning body. By Section 3.3 the angular momentum is fixed in space when no torque acts, while the angular velocity that the rotor law integrates is a different bivector and must wander. The torque-free Euler equation of the next section governs how.
3.8. Euler's equations without the cross product
The angular momentum of a torque-free body is fixed in space, while its angular velocity is a different bivector by Section 3.7. Hence the body-frame components of the angular velocity must change even when nothing acts on the body. The equations that govern that change are Euler's equations, and the matrix account writes them with a cross product of the angular velocity and the angular momentum. We derive them as a bivector rate law, a commutator, without the cross product of Section 1.9, and then treat the free top and its instabilities.
The derivation transports the conservation law into the body frame: differentiating the constant space angular momentum with the rotor law and stripping the sandwich rotors leaves a first-order equation for the body-frame angular momentum alone, whose natural product is the commutator of two bivectors.
The commutator product of two bivectors is half their commutator under the geometric product,
which is again a bivector in . It is the closed product on the space of bivectors and governs the rotational dynamics. It differs from the retired vector cross product of Section 1.9 though it shares the notation.
With this product the torque-free law takes one line.
For a rigid body under no torque, the body-frame angular-velocity bivector obeys
the rate of change of the body-frame angular momentum equal to its commutator product with the angular velocity. A nonzero body-frame torque adds to the right side.
Proof
In the space frame the angular momentum is constant under zero torque by [eq:dL-dt-tau]. Differentiate it with the rotor law in the form and its reverse : the two sandwich derivatives contribute and , so . Stripping the rotors and recognising the commutator product [eq:commutator-product] leaves [eq:euler-bivector].
Written in the principal basis of [eq:principal-axes], with and the planes oriented as for a right-handed principal frame , the single bivector equation [eq:euler-bivector] resolves into its three grade-two components, and the commutator products of the principal eigen-bivectors reproduce the classical Euler equations exactly,
with no cross product in the derivation. The antisymmetric coupling that the matrix account gets from is here the commutator [eq:commutator-product] of the bivectors. It is defined in every dimension and needs no dual. Two scalars are constant along any solution of [eq:euler-scalar]: the kinetic energy of [eq:L-inertia], and the squared magnitude of the space angular momentum , the first because the commutator product is orthogonal to its arguments and the second because is fixed in space.[3]
Take the symmetric top with two equal principal moments . The third of Euler's equations [eq:euler-scalar] reads , so the spin about the symmetry axis is constant. The remaining two become, on writing the constant ,
a plane rotation of the pair at the constant angular rate . The body-frame angular velocity therefore keeps a fixed component along the symmetry axis and a component of fixed length that precesses steadily around it. To check the conserved quantities, . Hence both the energy and the momentum magnitude are constant, as the general law requires. The tip of traces a circle about the symmetry axis, the simplest instance of the cone described next.
The trajectory of the angular-velocity bivector is a pair of cones. In the body frame the tip of is confined to the intersection of the energy surface and the momentum surface , two ellipsoids. That intersection is a closed curve, the polhode. Seen in the space frame the same motion sweeps a second curve, the herpolhode, as the body rolls its polhode on the invariable plane perpendicular to the fixed . The figure traces both curves. The signs of the coefficients in [eq:euler-scalar] decide the stability.
Order the moments and examine rotation nearly about each principal axis in turn, perturbing the two small components while the third stays near its steady value. About the largest axis set and small. Differentiating the second Euler equation of [eq:euler-scalar] and substituting the third gives
whose coefficient is negative, since and . Hence oscillates and the axis is stable. The identical computation about the smallest axis, with steady, again gives a negative coefficient and stability. About the middle axis, with steady, the two factors are and , the coefficient is positive, and the perturbation grows exponentially: rotation about the intermediate axis is unstable. This is the tennis-racket theorem, the flip of a spun book or racket about its middle axis. It follows from the sign of a single product in [eq:euler-scalar], the commutator [eq:commutator-product] that replaced the cross product.
3.9. Simple harmonic oscillator
The rigid body of the last three sections turned freely, its only law the conservation of the angular momentum plane. Tether a body to a fixed centre by a force that grows in exact proportion to the displacement, and the motion becomes the simple harmonic oscillation that underlies small vibrations of every system in this article, the pendulum of [pendulum-el] and the spinning body alike, since the leading term of any restoring force is linear. It is at once the simplest central force of Section 3.4 and in its anisotropic form the plainest use of the principal axes of Section 3.7. We take it in one, two and three dimensions. The conserved plane and the principal frame return.
A harmonic oscillator is a body of mass moving under a linear restoring force
where is a symmetric, positive-definite linear operator on , the stiffness operator, symmetric in the sense of the adjoint of [adjoint], . It is isotropic when is a positive scalar times the identity, so that is central in the sense of [eq:central-force] with , and anisotropic otherwise.
The second law [eq:second-law] turns [eq:sho-force] into the linear equation of motion . In one dimension the operator is a single positive number and the equation is the scalar one that names the phenomenon.
For with stiffness the equation of motion is
whose general solution is the sinusoid
with amplitude and phase fixed by the initial position and velocity, oscillating at the angular frequency independent of the amplitude. The energy
is conserved along the motion.
Proof
Dividing by gives , a linear second-order equation with constant coefficients whose characteristic roots are , so its real solutions are the span of and , which the amplitude-and-phase form [eq:sho-1d-sol] collects through . For the energy, differentiate [eq:sho-1d-energy] and substitute the equation of motion: , so is constant.
The amplitude-independence of the frequency in [eq:sho-1d-sol] is the property that makes the oscillator a clock, and belongs to the linear force alone: the pendulum of [pendulum-el], whose restoring force is linear only to first order, loses it once the swing widens. Two and three dimensions add the geometry the scalar case cannot show. Take first the isotropic oscillator, where the force is central and the conservation law of Section 3.4 applies unchanged.
For an isotropic oscillator in two or three dimensions, each Cartesian component obeys the one-dimensional equation [eq:sho-1d-eq] at the common frequency , so the trajectory is
for the constant vectors and . The angular momentum bivector [eq:angular-momentum] is the constant
and the orbit is an ellipse centred on the force centre, spanning the plane of , with and as a pair of conjugate semi-diameters.
Proof
Since , the equation separates componentwise, each component a copy of [eq:sho-1d-eq] at the same . Solving each in the cosine-and-sine basis and collecting the constant coefficient vectors gives [eq:sho-2d-sol], with and read off at . Differentiating, , and the areal bivector is
the cross terms collapsing through and , exactly as in the circular orbit [eq:circle-result]. Multiplying by gives the constant [eq:sho-2d-L]. Hence is conserved. Equation [eq:sho-2d-sol] is the parametric form of an ellipse centred at the origin with conjugate semi-diameters and , the image of the unit circle under the linear map that sends the axes to and .
Gravity's inverse-square attraction sends the body round an ellipse with the centre of force at a focus. The linear attraction sends it round an ellipse with the centre of force at the centre. Both conserve the same bivector [eq:angular-momentum], and the isotropic three-dimensional oscillator is no harder than the plane one, because the confinement of [central-conservation] already confines the motion to the fixed plane of , where [eq:sho-2d-sol] stands verbatim.
Anisotropy breaks the central symmetry and with it the conservation of the plane. Here the operator language of Part I applies. A symmetric stiffness has by the eigenblade proposition of [eigenblade] an orthonormal basis of eigenvectors, its principal axes, as does the inertia operator of [inertia-operator]. Along them the motion separates.
Let be a symmetric positive-definite stiffness operator with orthonormal eigenvectors and eigenvalues , so that . In the eigenbasis the equation of motion separates into three independent one-dimensional oscillators,
the normal modes, each vibrating at its own frequency along its principal axis. The general motion is their superposition, a closed curve when the frequencies are commensurable and a curve dense in a box when they are rationally independent.
Proof
Resolve the position in the eigenbasis, with , the components along the orthonormal principal axes of [eigenblade]. Because the are constant in the inertial frame, , while by linearity and the eigenvalue relation. The equation then reads , and matching the independent components gives . This is [eq:sho-3d-modes-eq]. Each is a one-dimensional oscillator solved by [eq:sho-1d-sol] at frequency , so . It closes into a periodic orbit exactly when the share a common period, that is when their ratios are rational. Otherwise it never repeats. When the only integer relation is the trivial one it covers densely the box ; a partial resonance such as with irrational confines it to a surface inside the box. These superpositions are the Lissajous figures.
The diagonalisation that separates the modes is the construction that gives the inertia operator of Section 3.7 its principal moments, and the eigenvalue problem the change of basis of [change-of-basis] set up in the matrix language: a symmetric operator resolved along the orthonormal frame in which it is diagonal. In the smallest nontrivial dynamics the oscillator combines the inner product, the adjoint, the eigenbasis and the conserved bivector of Part I, and every stable equilibrium imitates it to first order.
This closes the Newtonian and rigid-body dynamics. The position vector of Section 1.1 has acquired a mass and a law. The bivector of Section 1.6 has become the angular momentum and the angular velocity of a spinning body. The rotor of Section 1.10 has become the orientation itself, integrated by the singularity-free law [eq:rotor-equation], the third thesis anchor. The same position vector, tethered by a linear force, gives the harmonic oscillator, whose normal modes use the principal axes of the spinning body. Part IV recasts mechanics in the variational language of Lagrange and Hamilton in both tongues.
Part IV
4. Lagrangian and Hamiltonian mechanics
4.1. Configuration space and generalised coordinates
The three parts before this one followed a body through space by its position vector and a rigid body by its orientation rotor. Both are instances of one idea: a mechanical system at each instant is a point in a space of its possible arrangements. Mechanics studies the curve that point traces. This part recasts the dynamics of the earlier parts in that language, first in generalised coordinates and then on the rotor group itself in Section 4.3. We stay in coordinates throughout, as the Introduction promised, and treat the arrangement space as a coordinate patch.
The configuration space of a mechanical system is the set of its geometrically possible arrangements, and a set of generalised coordinates is a tuple
of real numbers that names each arrangement uniquely on a region of that set, the integer being the number of degrees of freedom. A motion of the system is a curve through the coordinates, and its generalised velocity is the tuple of time derivatives .
The generalised coordinates need not be lengths. A single particle in space is named by its three Cartesian components, so and ; a plane pendulum of fixed length is named by the one angle its rod makes with the vertical, so and ; the orientation of a rigid body about a fixed point is named by a rotor and has , the count the double cover of Section 1.10 fixed. A constraint reduces the coordinate count below the ambient one. This account admits the constraints that can be written as equations among the coordinates alone.
A holonomic constraint on a system whose arrangements sit in is a finite set of relations
among the positions and the time alone, each independent of the velocities. The relations cut the ambient coordinates down to independent ones, and any tuple [eq:gen-coords] of that many free parameters solving [eq:holonomic] is a valid set of generalised coordinates.
The pendulum bob moves in the plane, two Cartesian coordinates, subject to the one holonomic relation that its distance from the pivot is fixed. This leaves free coordinate, the angle. A bead on a rigid wire, a mass on a rigid rod and a body whose points keep fixed mutual distances are all holonomic systems, and the rigidity of Section 3.6 is itself the holonomic constraint that froze the distances there. Writing the physics in the free coordinates [eq:gen-coords] builds the constraint in from the start. No constraint force need ever be named, since the coordinates parametrise no other arrangements.
At each configuration the generalised velocity ranges over a copy of , the space of rates a curve through may have. The pair ranges over the numbers on which the Lagrangian of the next section is a function. The collection of all such pairs is the tangent bundle of the configuration space. Here it is the domain of coordinates and velocities on which mechanics is written.
4.2. Action and Euler-Lagrange equations
With the configuration named by the coordinates of Section 4.1, the law of motion is a condition on the whole path at once. A single scalar is assigned to every candidate curve between two fixed configurations and the system follows the curve that makes that scalar stationary.[4] This section defines the scalar, states the principle and derives the differential equations the true path obeys.
The Lagrangian of a system is a scalar function of the generalised coordinates [eq:gen-coords], their velocities and possibly the time. For a system of particles it is the kinetic energy less the potential energy, . The action of a path run between times and is the integral of the Lagrangian along it,
a real number assigned to the whole curve. Thus is a functional on the space of paths with fixed endpoints.
The action of [eq:action] takes a curve and returns a number, and the variational principle asks that number to be stationary. To make this precise compare the true path with a neighbouring one, , where is any smooth deformation that vanishes at the two ends, . Both curves then join the same endpoints. The action becomes an ordinary function of the number and stationarity is the vanishing of its derivative there.
A path with fixed endpoints is a physical motion when the action [eq:action] is stationary on it, meaning that for every deformation vanishing at and the first variation vanishes,
To turn the principle [eq:stationary] into equations, expand the derivative under the integral. Differentiating the action [eq:action] with respect to and setting , the chain rule applied to gives, summing over the coordinate index ,
The second term has the derivative of the deformation where the first has the deformation itself. Integrating it by parts moves the time derivative off ,
The boundary term of [eq:by-parts] vanishes because at the fixed endpoints. Substituting back into [eq:first-variation] collects the first variation onto with no derivative left on it,
The principle asks this to vanish for every admissible . The last step uses the lemma that turns an integral condition into a pointwise one.
If a continuous function on satisfies for every smooth vanishing at the endpoints, then identically on the interval.
Proof
Suppose is nonzero at an interior point. By continuity it keeps one sign on a small interval about it. Choosing a positive bump supported there and zero elsewhere makes the integral nonzero, against the hypothesis. So vanishes on the open interval and by continuity at the endpoints too.
The lemma turns the integral condition [eq:variation-collected] into the equation of motion.
A path is stationary for the action [eq:action] under fixed endpoints if and only if it satisfies, for each generalised coordinate,
These are the Euler-Lagrange equations, one second-order differential equation per degree of freedom.
Proof
The first variation was reduced in [eq:variation-collected] to a single integral of the bracketed expression against the deformation , with no derivative left on . Since the deformations of the several coordinates are independent, the fundamental lemma [fundamental-lemma] applied to each coefficient forces every bracket to vanish. This is [eq:euler-lagrange-eq]. Conversely, if the equations are satisfied the integrand of [eq:variation-collected] is zero and the action is stationary.
The equations [eq:euler-lagrange-eq] reproduce Newton at once for a particle in a potential. Take Cartesian coordinates with . Then , whose time derivative is , while . Thus [eq:euler-lagrange-eq] reads , the second law [eq:second-law] with the force the negative gradient of the potential, the gradient of Section 2.2. The same [eq:euler-lagrange-eq] holds in any coordinates, since the derivation never chose a frame, and the constrained systems of Section 4.1 are handled in their own free coordinates with no constraint force. Consider the pendulum.
A bob of mass swings on a rigid rod of length in a vertical plane under gravity , its single coordinate the angle from the downward vertical. The bob position is . Its speed is and its height is . The kinetic and potential energies give the Lagrangian
Compute the two pieces of [eq:euler-lagrange-eq] one at a time. The velocity derivative is , whose time derivative is . The coordinate derivative is . Thus the Euler-Lagrange equation is , that is
For small angles this is , simple harmonic motion of angular frequency and period . The rigidity of the rod never appears as a force since the coordinate builds it in.
4.3. Geometric-algebra Lagrangian mechanics
The coordinates of Section 4.1 serve a particle, but the configuration of a rigid body is a rotor, an element of the algebra, and coordinatising it by angles is the move that gimbal-locked in Section 3.6. In geometric algebra we vary the rotor directly as a multivector. This needs a derivative with respect to a multivector variable. This section builds it, states the Euler-Lagrange equation for a rotor-valued configuration and recovers the Euler equations of Section 3.8 from a single Lagrangian on the group .
The scalar product of two multivectors, the scalar part of their geometric product, pairs the algebra with itself and suffices to differentiate a scalar function of a multivector.
Let be a scalar-valued function of a multivector variable . Its multivector derivative is the multivector defined by its scalar product against every direction ,
the right side being the ordinary directional derivative of along . In a blade basis with reciprocal blades it reads , the gradient in the coordinates of the algebra. It respects grade like the vector derivative of Section 2.2 that it generalises.
A Lagrangian for a rotor-valued configuration is a scalar function of the rotor and its rate, and Hamilton's principle [eq:stationary] applies verbatim, the deformation now an even-grade vanishing at the ends. Repeating the variation of Section 4.2 with the multivector derivative in place of the partial derivatives, and integrating by parts against the scalar product, gives the Euler-Lagrange equation in the same shape as before.
A path in the even subalgebra is stationary for the action of a Lagrangian under fixed endpoints and unconstrained even-grade deformations if and only if
with and the multivector derivatives [eq:mv-derivative] with respect to the rotor and its rate. A rotor is constrained to by , and [eq:rotor-el-eq] alone does not impose it: the admissible deformations must stay tangent to the group, or the constraint enters through a multiplier. The cleanest route to the dynamics writes the deformations as with a bivector and reduces to the body-frame angular velocity. Deformations of that form keep to first order.
Proof
The variation of Section 4.2 carries over with the multivector derivative [eq:mv-derivative] in place of the partial derivatives and the scalar product in place of the coordinate sum. Taking the first variation of the action of along an even-grade deformation that vanishes at the ends, integrating the term by parts against the scalar product, and dropping the boundary term by the fixed endpoints as in [eq:by-parts], collects the variation onto against . Stationarity for every gives [eq:rotor-el-eq].
The reduction recovers the free top. The kinetic energy of a body turning about a fixed point is the scalar of Section 3.7, built from the body-frame angular-velocity bivector of Section 3.5, and with no potential the Lagrangian is that energy alone,
Vary the rotor path and let be the bivector that measures the deformation in the body frame, the variational counterpart of as is of . Both and are pure bivectors, since the rotor lives on the unit sphere of the even subalgebra of Section 1.10 and its tangent directions are grade two. Differentiating the definitions and using the reverse of a bivector being its negative, the variation of the body angular velocity and the time derivative of the generator are related by
the commutator product [eq:commutator-product] of the bivectors appearing exactly as it did in the Euler law. This is the one identity the reduction turns on. It states that the body-frame velocity and its variation do not commute because the rotor group does not.
The Lagrangian [eq:spin3-lagrangian], made stationary over rotor paths with fixed endpoints, yields the torque-free Euler equation of Section 3.8,
identical to [eq:euler-bivector] and resolving in the principal basis into the scalar equations [eq:euler-scalar].
Proof
The derivation is the variation of Section 4.2 carried through in bivectors. Because the inertia operator is symmetric under the bivector scalar product, the first variation of the action is
using [eq:delta-omega] for . Integrate the term by parts, the boundary term dying because vanishes at the ends as in [eq:by-parts], and rewrite the commutator term with the invariance of the scalar product under the commutator, , the bivector form of the cyclic symmetry of the scalar triple product. Both steps move everything onto ,
Stationarity for every bivector deformation , by the fundamental lemma [fundamental-lemma] read on the bivector-valued integrand, forces the bracket to vanish, and since that is exactly [eq:euler-poincare-eq].
This recovers the rigid body of Part III from a variational principle with the orientation a rotor throughout. No Euler angle enters and no coordinate singularity arises. The commutator product that replaced the cross product in the equations of motion emerges from the geometry of the variation.
4.4. Noether's theorem
The conservation laws that Part III proved one at a time, momentum from the absence of a force and the angular momentum plane from the absence of a torque, are instances of a single theorem. Every continuous symmetry of the action gives a quantity that stays fixed along the motion and is read off the symmetry by one formula. This section states and proves that correspondence, then reads energy and the angular momentum bivector of Section 3.3 off the symmetries that produce them.
A continuous symmetry is a family of deformations of the coordinates, indexed by a parameter , that leaves the Lagrangian unchanged. Write the infinitesimal deformation as , the generator of the family at .
A vector field on the configuration space generates a continuous symmetry of the Lagrangian when the infinitesimal change , together with the induced , leaves the Lagrangian invariant,
Combined with the equation of motion the invariance [eq:symmetry-invariance] makes the change in the Lagrangian a total time derivative of the conserved charge.
To every continuous symmetry [eq:symmetry-invariance] of the Lagrangian there corresponds a conserved quantity, the Noether charge,
constant along every physical motion. If the deformation changes the Lagrangian by a total time derivative , the conserved charge is .
Proof
On a physical path the coordinate derivative equals by the Euler-Lagrange equation [eq:euler-lagrange-eq]. Substituting it into the invariance [eq:symmetry-invariance] and recognising the product rule,
the expression is a total time derivative. Thus the charge [eq:noether-charge] is constant along the motion. When the deformation changes the Lagrangian by , the same step leaves .
The angular momentum of Section 3.3 is the charge of rotational symmetry, and geometric algebra states it as the whole bivector. Let the Lagrangian of a particle in a central potential be turned by the rotor generated by a fixed unit bivector . The infinitesimal rotation is the contraction of Section 3.5. Since it preserves both the speed and the distance from the centre, [eq:symmetry-invariance] holds. The Noether charge [eq:noether-charge] is the momentum paired with the generator,
the scalar product of the generating plane with the angular momentum bivector of [eq:angular-momentum]. Since this is conserved for every choice of the plane , the whole bivector is conserved. This is the constancy of the angular momentum plane that Section 3.4 derived from the vanishing torque, now a consequence of rotational symmetry. Where the matrix account conserves the three components of an axial vector, Noether's theorem gives the oriented plane directly.
The other cardinal example is energy, the charge of symmetry under translation in time. When the Lagrangian has no explicit , shifting the whole motion earlier or later leaves the action unchanged, and the corresponding conserved quantity is built from the velocities and the Lagrangian.
When the Lagrangian has no explicit time dependence, , the energy
is conserved along every motion. For with quadratic in the velocities, , the total mechanical energy.
Proof
That [eq:energy] is constant follows from the chain rule and the Euler-Lagrange equation. Differentiating along the motion gives . Replacing the first factor by collects the right side into . Hence .
The combination becomes the Hamiltonian in the next section.
4.5. Legendre transform and Hamiltonian
The Euler-Lagrange equations of Section 4.2 are second-order equations in the coordinates and velocities. A second formulation, equivalent in content, trades each velocity for a momentum and replaces the second-order equations by first-order ones on a doubled space. The passage between the two is a Legendre transform, and its output is the Hamiltonian. This section performs the transform and states Hamilton's equations, then works the pendulum in the new variables.
The quantity conjugate to a coordinate is the velocity derivative of the Lagrangian that already appeared inside the Euler-Lagrange equation and the Noether charge.
The conjugate momentum to the coordinate is
the same object whose time derivative is the left side of the Euler-Lagrange equation [eq:euler-lagrange-eq]. For a free particle is the ordinary momentum of [eq:momentum]; in general it is a different combination with the index of the coordinate to which it is conjugate.
The Legendre transform takes the Lagrangian on coordinates and velocities to a function on coordinates and momenta by exchanging the velocity for the momentum as the independent variable. The transform is well defined when the momenta [eq:conjugate-momentum] can be solved for the velocities. This holds when the Lagrangian is convex in the velocities, as a kinetic energy quadratic with positive mass always is.
The Hamiltonian is the Legendre transform of the Lagrangian in the velocities,
with every velocity on the right expressed through [eq:conjugate-momentum] as a function of the coordinates and momenta. The variables are the coordinates of phase space, and a state of the system is a single point there.
The Hamiltonian [eq:hamiltonian-def] is the conserved energy [eq:energy] of Section 4.4, now a function on phase space. The equations of motion in the new variables follow by differentiating [eq:hamiltonian-def] and using the definition of the momentum to cancel the velocity terms. The result is a symmetric pair.
The Euler-Lagrange equations [eq:euler-lagrange-eq] are equivalent to the first-order Hamilton equations
one pair per degree of freedom for the motion of the phase-space point.
Proof
Take the differential of [eq:hamiltonian-def]. The velocity terms cancel by the definition [eq:conjugate-momentum] of the momentum: loses its terms and leaves . Reading off the coefficients gives and by the Euler-Lagrange equation. These are the two equations of [eq:hamilton-eqs-pair]. Conversely, the first of them inverts the Legendre transform and recovers , and the second then reads .
The single second-order law has become two first-order laws, symmetric under the near-exchange of and that the next section makes exact.
Return to the pendulum Lagrangian [eq:pendulum-lagrangian], . Its conjugate momentum [eq:conjugate-momentum] is , which inverts to . The Hamiltonian [eq:hamiltonian-def] is
the kinetic term now a function of the momentum and the potential unchanged. Hamilton's equations [eq:hamilton-eqs-pair] read
To check that these reproduce the Lagrangian result, differentiate the first and substitute the second: , exactly the pendulum equation [eq:pendulum-eom]. The two first-order equations of [eq:pendulum-hamilton-eqs] describe the same swing as the one second-order equation, as a flow of the phase-space point whose portrait the next section animates.
4.6. Symplectic bivector and Poisson brackets
Hamilton's equations of Section 4.5 are already nearly symmetric between coordinate and momentum, and the last step exposes the structure that makes the symmetry exact. Phase space has a distinguished bivector, and against it every observable acquires a rate of change written as a single bracket. This section introduces that bivector, defines the Poisson bracket it induces, casts Hamilton's equations in bracket form, and reads off the conservation of phase-space volume that the bracket structure guarantees.[5]
The pairing of each coordinate with its conjugate momentum is an oriented plane in phase space, and summing those planes over the degrees of freedom gives one bivector in the sense of the grade-two objects of Part I, now built on the -dimensional phase space.
On the phase space with coordinates , the symplectic bivector is the sum of the coordinate-momentum planes,
a fixed grade-two element pairing each coordinate direction with its conjugate momentum direction. It is nondegenerate: the only phase-space direction whose contraction with vanishes is the zero direction, and it equips phase space with the oriented-area measure that the dynamics preserves.
The symplectic bivector turns the gradient of any phase-space function into a flow, and applying that to two functions in turn gives their Poisson bracket, the antisymmetric pairing of the Hamiltonian formalism.
The Poisson bracket of two functions and on phase space is
the contraction of their gradients through the symplectic bivector [eq:symplectic-bivector-def]. It is bilinear and antisymmetric, , obeys the Jacobi identity, and on the coordinates themselves takes the canonical values
The fundamental bracket of [eq:canonical-brackets] is the phase-space statement that a coordinate and its own momentum are conjugate. It is the one relation from which the whole bracket algebra is built. Hamilton's equations [eq:hamilton-eqs-pair] now collapse to a single form that governs every observable at once.
Any phase-space function with no explicit time dependence evolves along the motion by its bracket with the Hamiltonian,
and Hamilton's equations [eq:hamilton-eqs-pair] are the special cases and . A quantity is conserved exactly when its bracket with the Hamiltonian vanishes, and by antisymmetry recovers the conservation of energy of Section 4.4.
Proof
Equation [eq:bracket-motion] is the chain rule read through the bracket: , and substituting Hamilton's equations [eq:hamilton-eqs-pair] for and turns the right side into the bracket [eq:poisson-bracket-def] of with .
That a conserved quantity is one commuting with the Hamiltonian under the bracket is the Hamiltonian face of Noether's theorem, and the two examples of Section 4.4 reappear as for the energy and for the angular momentum plane in a central field.
The flow generated by the Hamiltonian preserves the symplectic bivector [eq:symplectic-bivector-def], and with it the oriented-volume measure that the bivector defines. This is Liouville's theorem.
The phase-space flow of Hamilton's equations [eq:hamilton-eqs-pair] preserves the symplectic bivector [eq:symplectic-bivector-def], and therefore the phase-space volume it induces: a region of initial conditions carried along by the flow keeps its -dimensional volume for all time, though its shape may distort without limit. Equivalently the phase-space velocity field is divergence-free,
the two second derivatives cancelling since the mixed partials commute.
Proof
The computation [eq:liouville-div] proves the volume statement: the flow of Hamilton's equations has no sources or sinks and moves phase-space volume as an incompressible fluid does. For the bivector itself write , so that Hamilton's equations read with . The derivative of the flow map obeys with the symmetric Hessian of , and . So for all time. That identity is the preservation of [eq:symplectic-bivector-def].
For a single degree of freedom the volume is an oriented area, and any patch of the phase plane keeps its area as it is swept along however the Hamiltonian shears it. The pendulum makes it visible: a ring of initial conditions in the plane of [eq:pendulum-hamilton-eqs] circulates and stretches under the flow while the area it encloses stays constant.
The closed contours of the figure are the level sets of the Hamiltonian [eq:pendulum-hamiltonian], each a motion of fixed energy, the small ones near the bottom the gentle swings and the large ones the rotations over the top, with the dividing separatrix the swing that reaches the upright. The advected patch keeps its oriented area exactly. This is the one-degree-of-freedom case of Liouville's theorem [liouville], and the symplectic bivector [eq:symplectic-bivector-def] is the invariant the flow respects. The oriented plane that opened the article as the outer product of two vectors returns as the symplectic bivector on phase space.
4.7. Scale-graded norms on jets and multivectors
Two graded towers have run the length of the article: the jet of Section 1.1, which stacks a trajectory by order of differentiation, and the multivector of Section 1.7, which stacks the algebra by grade. Section 1.1 promised that a single operator-norm estimate, repeated up the jet, would govern every approximation. This section discharges the promise by giving both towers one norm, weighted by the scale at which each level acts.
Evaluate the -jet of a map on a displacement of size . The -th level is the -th Fréchet derivative , a symmetric -linear map, and its contribution to the Taylor expansion of [eq:increment-bound] is , whose size is at most . The -th level enters at order , and each rung of the jet is weighted by the power of the displacement it consumes. Summing the rungs gives the size of the jet at scale .
The scale-graded norm of the -jet of at displacement scale weights the -th derivative by the order at which it acts,
As it returns the value alone; as grows the higher derivatives take over in order, and the whole sum bounds the size of the degree- Taylor polynomial over a step of size .
The definition makes the tower of Section 1.1 quantitative. A bound on level controls the variation of level across a step, and [eq:scale-graded-jet-norm] collects those controls into one number in which each level is weighted by the power of the scale it works at. Dividing by isolates the top level as , and letting isolates the value.
The same weighting fits the algebra, because a grade- blade is an -dimensional volume element and scales as the -th power under a dilation of space. Assigning the grade- slot the weight is the algebraic image of weighting the -th derivative at order , and the grade decomposition of [eq:grade-sum] then has a scale-graded norm of its own.
For a multivector in and a scale ,
where is a fixed norm on the grade- subspace . At it reduces to the grade-blind sum .
The pair [eq:scale-graded-jet-norm] and [eq:scale-graded-mv-norm] is the parity of Section 1.1 made metric: derivative order and grade are one kind of index, each weighted by raised to it. The construction separates two considerations that the bare sum runs together. The grade-specific part is the per-slot size , measured within one grade where the norm is an ordinary length; the grade-free part is their weighted assembly, in which the scale sets the relative weight of the slots. Setting forgets the grading and reads the whole element as one. Letting vary restores it.
The standard function-space norms grade the same tower. Let be a domain and write for [eq:scale-graded-jet-norm] with every derivative taken at . For fixed the supremum of over is equivalent to the norm : the supremum of a sum of non-negative terms lies between the largest of their suprema and the sum of their suprema, and the weights are fixed positive constants. The Hölder space , with , keeps that norm and adds the seminorm . A bound on level would control the variation of level across a step of size by a multiple of ; the Hölder condition asks for a multiple of instead and needs no level at all. Replacing the supremum over by the norm of , with the derivatives now taken in the weak sense, gives a norm equivalent to the Sobolev norm of by Minkowski's inequality and the same comparison of a sum with its largest term. The equivalence constants depend on , and . The estimates of this article use only the case.
4.8. Coda on the two languages
The construction is complete. From an inertial frame and the identification of a point with its position vector the article built the algebra of a single instant, the calculus that differentiates it, the Newtonian and rigid-body dynamics it governs and the variational formalism that reorganises the whole. It did so in two languages kept in step, and the second language was tested at three anchors.
The first anchor was the catalogue of discomforts of the matrix language in Section 1.5. A rotation of the plane had no real eigenvalue, the axis of a spatial rotation was an accident of odd dimension, the cross product was a three-dimensional pseudovector that reflected with the wrong sign, and the determinant and orientation were bolted on from outside the algebra of vectors. The geometric product retired every one of them. The imaginary unit that the plane rotation demanded became the unit bivector squaring to of Section 1.8, already inside the real algebra of space; the axis gave way to the plane of rotation, defined in every dimension; the cross product was unmasked in Section 1.9 as the disguised dual of the bivector ; and the determinant became, in Section 1.11, the scale by which an operator inflates the pseudoscalar within the algebra.
The second anchor was the calculus, tallied in Section 2.5. The gradient, divergence and curl, three operators of incompatible type with the curl confined to three dimensions, became the grade parts of one vector derivative ; and the four integral theorems of vector analysis, each welded to its dimension and its grade, became grade projections of the single fundamental theorem of geometric calculus. One operator and one theorem, defined in every dimension and needing no handedness convention, stood where the classical apparatus spread a zoo.
The third anchor was the dynamics, the rotor kinematics of Section 3.6. Orientation given by Euler angles gimbal-locks, its equations dividing by a sine that vanishes when two axes align, and rotation matrices have nine numbers under six constraints. The rotor law has no coordinates on the orientation and hence no coordinate singularity; its exact flow preserves the normalisation, a numerical step restores it by one division, and rotors compose by multiplication. The same rotor returned in Section 4.3 as the configuration of a variational principle, and the Euler equations followed from the geometry of the variation with the commutator product in place of the cross product.
The second language contributed a single change of vantage kept up for the length of the article: the oriented plane in place of the axis, the whole element of the algebra in place of its shadow in coordinates. The bivector that first appeared as the outer product of two vectors became the angular momentum of an orbit, the angular velocity of a spinning body, the generator of a rotor and finally the symplectic form on phase space. It was one object throughout, and geometric algebra follows it through each appearance. The matrix language kept every formula computable; the geometric language kept every formula meaning what it said. Classical mechanics, constructed from an inertial frame and nothing else, is legible in both, and most legible where the two are read together.
References
- D. Hestenes, G. Sobczyk (1984). Clifford Algebra to Geometric Calculus. Reidel. DOI
- C. Doran, A. Lasenby (2003). Geometric Algebra for Physicists. Cambridge University Press. DOI
- D. Hestenes (1999). New Foundations for Classical Mechanics, 2nd ed. Kluwer Academic Publishers. DOI
- V. I. Arnold (1989). Mathematical Methods of Classical Mechanics, 2nd ed. Springer-Verlag. DOI
- J. E. Marsden, T. S. Ratiu (1999). Introduction to Mechanics and Symmetry, 2nd ed. Springer. DOI