Primer On Geometric Algebra: For Introductory Mathematics and Physics
Primer On Geometric Algebra: For Introductory Mathematics and Physics
“Philosophy is written in that great book which ever lies before our eyes –– I
mean the universe –– but we cannot understand it if we do not first learn the
language and characters in which it is written. This language is mathematics,
and the characters are triangles, circles and other geometrical figures, without
whose help it is impossible to comprehend a single word of it; without which
one wanders in vain through a dark labyrinth.”
1
Primer on Geometric Algebra
OUTLINE
2
I. Prolog: On optimizing the design of introductory mathematics.
Physics teachers are universally dismayed by the paltry understanding of mathematics that
students bring from their mathematics courses. Blame is usually laid on faulty teaching. But I
hold that the crux of the problem is deeply embedded in the curriculum. From the perspective of
a practicing scientist, the mathematics taught in high school and college is fragmented, out of
date and inefficient!
The central problem is found in high school geometry. Many schools are dropping the course as
irrelevant. But that would be a terrible mistake for reasons already clear to Galileo at the dawn of
science.
• Geometry is the starting place for physical science, the foundation for mathematical
modeling in physics and engineering and for the science of measurement in the real
world.
• Synthetic methods employed in the standard geometry course are centuries out of date;
they are computationally and conceptually inferior to modern methods of analytic
geometry, so they are only of marginal interest in real world applications.
A basic pedagogical principle: The depth and extent of student learning is critically dependent
on the quality of the available mathematical tools.
Whether or not the high school geometry course can be reformed in practice, the course content
deserves to be reformed to make it more useful in applications.
Objective of this workshop: To demonstrate with specific examples how geometric algebra
unifies high school geometry with algebra and trigonometry and thereby simplifies and facilitates
applications to physics and engineering.
3
II. Standard algebraic tools for linear geometry: Vector Addition and scalar multiplication.
The term scalar refers to a real number or variable, with properties taken for granted here.
• Commutative
a+b=b+a
• Associative
(a + b) + c = a + (b + c)
a + (–a) = 0
a+0=a
• Subtraction
a – b = a + (–b)
4
B. Rules for multiplication by scalars (denoted by Greek letters and/or italics)
1a = a, (–1)a = –a, 0a = 0
• Distributive
α(a + b) = αa + αb
(α + β )a = αa + βa
a + a + a = (1 + 1 +1 )a = 3a
• Associative
α(βa) = (αβ)a
• Commutative
αa = aα
• Magnitude and direction: Every vector a has a unique scalar magnitude a = a and (if
a ≠ 0) a direction aˆ so that
a = aaˆ
• Collinearity. Nonzero vectors a and b are said to be collinear or linearly dependent if there
is a scalar β such that
b = βa
x(α1,α 2 , ,α n ) = α1a1 + α 2a 2 + + α na n
is not zero for any combination of scalars α1,α 2 , ,α n (not all zero). The scalars { α1,α 2 , ,α n }
are said to be coordinates for the vector x(α1,α 2 , ,α n ) with respect to the basis { a1,a 2 , ,a n }.
The set { x(α1,α 2 , ,α n )} for all values of the coordinates is an n-dimensional vector space.
5
C. Parametric equations
Exercise: As appropriate, identify and sketch the indicated geometric figures below.
When is it necessary to designate a particular point by the zero vector?
• Line: x(α) = αa + b
Line segment for 0 ≤ α ≤ 1
• Plane: x(α, β) = αa + βb + c
1. α 2 + β 2 = 1
2. β = α 2
3. α 2 − β 2 = 1
Solutions
1. x(θ ) = a cosθ + bsin θ + c
2. x(α ) = aα + bα 2 + c
3. x(θ ) = a coshθ + bsinh θ + c
News Release: Physics Education Research in a large state university found that, after
completing a semester of introductory physics, most students were unable to carry out graphical
vector addition in two dimensions. The more complex skills of coordinating scalar multiplication
with vector addition were not investigated. [Nguyen & Meltzer, AJP 71: 630-638 (2003)]
Question: What are likely reasons for this unacceptable failure of mathematics instruction?
Answers:
• Failure of the math curriculum to provide timely instruction in vector methods.
• Over reliance on coordinate methods in most courses.
• Vectors are only sporadically employed and usually with orthogonal bases, so students
have little opportunity to develop fluency with the general features of vector algebra
listed above.
• Students are unclear about the geometric interpretation of vectors (see below)
• Vector algebra is incomplete without rules for multiplying vectors that encode
information about magnitudes and relative direction (see GA below).
6
PART I. Introduction to Geometric Algebra and Basic Applications
The algebraic properties of vector addition and scalar multiplication are insufficient to
characterize the geometric concept of a vector as a directed line segment, because they fail to
encode the properties of magnitude and relative direction. That deficiency is corrected by
defining suitable algebraic rules for multiplying vectors.
7
We have shown that the geometric product interrelates three kinds of algebraic entities: scalars
(0-vectors), vectors (1-vectors), and bivectors (2-vectors) that can be interpreted as geometric
objects of different dimension. Geometrically, scalars represent 0-dimensional objects, because
they have magnitude and orientation (sign) but no direction. Vectors represent directed line
segments, which are1-dimensional objects. Bivectors represent directed plane segments, which
are 2-dimensional objects. It may be better to refer to a bivector B = BB̂ as a directed area,
because its magnitude B = B is the ordinary area of the plane segment and its direction B̂
represents the plane in which the segment lies, just as a unit vector represents the direction of a
line. The shape of the plane segment is not represented by any feature of B, as expressed in the
following equivalent geometric depictions (with clockwise orientation):
B = = B = B
Given any non-zero vector a in the plane of bivector B, one can find a vector b such that
B = ba = –ab,
B2 = − B = − a 2 b 2 ,
2
8
IV. Rotors and rotations in the Euclidean plane
Let i denote the unit bivector for the plane. As proved above, i 2 = −1, so i has the properties of
the unit imaginary in the complex number system. However, by relating it multiplicatively to
verctors, GA endows i with two new geometric interpretations:
ia = –ai = b,
i = ba −1 (counterclockwise sense)
The operator interpretation of i generalizes to the concept of rotor Uθ , the entity produced by
the product ba of unit vectors with relative direction θ.
a 2 = b2 = 1, i 2 = −1
b ⋅ a ≡ cosθ
b ∧ a ≡ isinθ
Rotor:
Uθ = ba = b ⋅a + b ∧ a = cosθ + i sinθ ≡ e iθ
9
Rotor equivalence of directed arcs is like
z = λU = λ e iθ = ba
z † = λU † = λ e − iθ = ab
zz † = λ 2 = (ba)(ab) = a 2b 2 = z
2
Modulus: z = λ = a b
10
Relation to standard complex number notation: z = Re z + i Im z = ba
with Re z = 12 ( z + z † ) = b ⋅ a i Im z = 12 ( z − z † ) = b ∧ a
(Note: this representation of complex numbers in a real GA is a special case of spinors for 3d).
Exercise: Prove the following vector identities and show that they are equivalent to trig
identities for unit vectors in a common plane.
Exercise: Use rotor products to derive the trigonometric double angle formulas:
(c) cos(θ + ϕ ) = ??
(b) sin(θ + ϕ ) = ??
11
Laws of the triangle follow directly from the geometric product:
a 2 + b 2 − 2 ab cos γ = c 2
News Bulletin: The World Health Organization has announced a world-wide epidemic of the
Coordinate Virus in mathematics and physics courses at all grade levels. Students infected with
the virus exhibit compulsive vector avoidance behavior, unable to conceive of a vector except as
a list of numbers, and seizing every opportunity to replace vectors by coordinates. At least two
thirds of physics graduate students are severely infected by the virus, and half of those may be
permanently damaged so they will never recover. The most promising treatment is a strong dose
of Geometric Algebra.
It may be surprising that the concept of vector is so difficult for students, since intuitive notions
of direction and distance are essential for navigating the everyday world. Surely these intuitions
need to be engaged in learning the algebraic concept of vector, as they are essential for
applications. The necessary engagement occurs only haphazardly in conventional instruction,
and that is evidently insufficient for most students.
One barrier to developing the vector concept is the fact that the correspondence between vector
and directed line segment has many different interpretations in modeling properties of real
objects and their motions:
12
• Vectors as points designate places in a Euclidean space or with respect to a physical
reference frame. Requires designation of a distinguished point (the origin) by the zero
vector.
• Position vector x for a particle which can “move” along a particle trajectory x = x(t)
must be distinguished from places which remain fixed.
• Kinematic vectors, such as velocity v = v(t) and acceleration are “tied” to particle
position x(t). Actually, they are vector fields defined along the whole trajectory.
• Dynamic vectors such as momentum and force representing particle interactions.
• Rigid bodies. It is often convenient to use a vector a as a 1d geometric model for a rigid
body like a rod or a ruler. Its magnitude a = a is then equal to the length of the body, and
its direction aˆ represents the body’s orientation or, better, its attitude in space. The
endpoints of a correspond to ends of the rigid body, as expressed in the following
equation
x(α ) = x 0 + αa {0 ≤ α ≤ 1}
for the position vectors of a continuous distribution of particles in the body. Note the
crucial distinction between curves (and their parametric equations) that represent particle
paths and curves that represent geometric features of physical bodies.
(1) Discuss and sketch values of the parametric equation x = αa + βb with the
constraint α + β = 1.
( x − a) ∧ u = 0 .
Exercises:
Hints:
13
(3) x ∧ u = a ∧ u = du. Sketch the directed areas and solve for d.
Rigid displacements: Congruence and measurement
x → x′ = Rx + a R = e iθ
y → y′ = Ry + a Ry = e −iθ
r = x − y → r ′ = x ′ − y′ = R(x − y′ ) or r ′ = Rr = r Ry
x(t ) = x 0 + vt
[Note that including the origin or coordinates in the map introduces arbitrary and
unnecessary complications.]
Derive a nonparametric equation for this model, and relate it to angular momentum.
14
B. Constant acceleration model.
(without coordinates!!)
dv
=g v = v 0 + gt
dt
(hodograph)
dx
=v x = x 0 + v 0 t + 12 gt 2
dt
(trajectory)
r
1
2 (v + v 0) = ≡v
t
v − v 0 = gt
“solving a parallelogram!!”
Problem: Determine
(a) the range r of a target sighted in a direction rˆ that has been hit by a projectile
launched with velocity v0 ;
(b) launching angle for maximum range;
(c) time of flight
15
Solving a parallogram with GA:
v0 2r/t
2r
v − v 0 = gt v + v0 =
t
gt
We eliminate t by multiplication, to get:
v
(v − v 0 )(v + v 0 ) = 2rg
Expanding, we obtain
v 2 − v0 2 + vv 0 − v 0 v = 2 ( r ⋅ g + r ∧ g )
2 v ∧ v0
v 2 − v0 2 = 2 r ⋅ g v ∧ v0 = r ∧ g
r ⋅g = 0 v02 = v 2 . v0
So v ∧ v0 = r ∧ g = rg θ 2r/t
θ
gives v02 sin 2θ = rg gt
v
v2
Hence the range formula: r = 0 sin 2θ
g
The general GA solution for an elevated target and time of flight is derived in NFCM.
16
C. Circular motion. Motion on a circle of radius r = r centered at point c:
dr
Velocity: v= = Ωr = ωir
dt
dθ
Angle: θ > 0 for counterclockwise motion Angular speed: ω=
dt
dR
Rotor eqn. of motion: = ΩR
dt
UCM θ = ωt r = e iω tr0
x = x(s) = parametric eqn. for particle path. [Sketch general motion map]
ds dx iθ
Speed: v = Tangent direction: ˆv = = e vˆ 0
dt ds
dv
Exercise: A particle slides on the frictionless track below subject to a force mg + N = m
dt
that keeps it on the track. Sketch its velocity v and acceleration a at the indicated points:
17
D. General Keplerian motion (under an inverse square force). Full treatment in NFCM
dv k
(1) Dynamical model: m = − 2 rˆ
dt r
2 drˆ
(a) Angular momentum: L = mr ∧ v = mr rˆ
dt
dv d
Hints: (a) mr ∧ = ( mr ∧ v) = r ∧ f(r) = 0 (Central force)
dt dt
dv kL drˆ
(b) L = − 2 rˆ = k
dt mr dt
(3) Model analysis. Derive the following algebraic features of the model:
[Details are fully worked out in NFCM]
2 k
(a) Energy E = 12 m v − is constant and related to eccentricity by
r
2
2L E
ε 2 − 1=
mk 2
18
k
(c) Hodograph v = v(rˆ ) = (rˆ + ε ) (orbit in velocity space)
L
k2
Show this is a circle by deriving the non-parametric eqn. (v − u)2 = .
L2
(e) Initial value problem: Determine L and ε given one value of v and rˆ .
(4) Scattering problem: Given asymptotic initial velocity and angular momentum, find v f = v final
Lv 0 + k
vf = v This solves the scattering problem completely!!
Lv 0 − k 0
19
Scattering angle Θ and impact parameterb defined as follows:
v f = v 0e i Θ = e −i Θ v 0
2bE
L = bmv0 = Lˆ = i (k > 0) or Lˆ = −i ( k < 0)
v0
Lv 0 − k 2Ebi − k k
e iΘ = = b= cot 12 Θ
Lv0 + k 2Ebi + k 2E
A reference system assigns each particle a definite position with respect to a given rigid body
(or reference frame). The set of all possible position vectors is a 3d Euclidean vector space
called the position space of the frame.
Time dependent rigid displacement of one reference frame with respect to another is
completely specified by a time dependent
dx dR
Problem: Suppressing the time argument and writing x= , R= = WR ,
dt dt
derive the following equations relating velocities and accelerations in the two frames aligned at
time t.:
x′ = x + Wx + a
x′ = x + 2Wx + (W + W 2 )x + a
Newton’s 1st Law (implicitly) defines an inertial frame as a rigid body with respect to which
every free particle has constant velocity.
Principle of Relativity requires that the laws of physics are the same in all inertial systems.
[First formulated by Galileo and incorporated by Newton as a corollary in his theory].
20
Problem: Apply this to Newton’s 2nd law to prove that any two inertial frames are related by a
Galilean transformation
Exercise: Inside a cable car climbing a slope with constant velocity v0 an object is dropped
from rest. Derive eqns. for the trajectory within the car and with respect to the earth outside.
Problem: Discuss invariance of Newton’s first and second laws with respect to Galilean time
translation and scaling: t → t′ = α t + β , where α and β are constants.
X. The Zeroth Law of physics defines the fundamental presumptions about space, time and
existence of real entities that underlie all of physics.
Newton’s First Law defines an inertial system and a uniform time scale.
The remaining four Laws of Newtonian physics define particle interactions and dynamics.
[Ref. Modeling Games in the Newtonian World, AJP 60: 732 (1992)]
21
PART II. Special Relativity with Geometric Algebra
[Ref. NFCM, 2nd Ed. (1999), Chap. 9]
XI. Defining Spacetime
1. Principle of Relativity. Einstein adopted this principle from Newtonian theory, but
raised its status from a mere corollary to a basic principle.
2. Invariance of the speed of light. Einstein assumed that the speed of light c has the same
value in all inertial systems.
Minkowsky (1908) incorporated these principles into a new conceptual fusion of space and time
that can be defined with GA by the following assumptions:
1. In a given inertial system, the time t and place x is of an event is represented as a single
point X = ct + x in a 4-dimensional space called spacetime (see spacetime maps below).
∆X = X 2 − X1 = c(t2 − t1 ) + (x 2 − x 2 ) = c∆ t + ∆x
between events X 2 and X1 has an invariant magnitude ∆X called the proper distance
between the events and given by
where ∆X˜ = c ∆t − ∆x , and the signature ε of the interval has the value 1, 0 or –1, and,
respectively, the interval is said to be timelike, lightlike, or spacelike.
∆X∆X˜ = ε ∆ X 2 = ε ∆ X ′ 2 = ∆ X ′∆ X˜ ′
wnere ∆X ′ = X ′2 − X1′ = c ∆t′ + ∆x ′ is the same interval represented in some other inertial
system.
Problem: Prove that spacetime interval invariance implies that the speed of light c has the same
constant value in all inertial systems.
22
XII. Spacetime Maps
Spacetime maps are essential for physical interpretation of GA equations in relativity theory.
This is a spacetime map showing the lightcone for an event X 0 . Events X k (k = 1, 2, 3) lie on
straight lines passing through X 0 . Intervals ∆X k = X k − X 0 are said to be timelike, lightlike or
spacelike, respectively, as they lie inside, on, or outside the invariant lightcone.
23
This is a spacetime map of events in a timelike plane showing the position vector x = X 2 − X1
for the event X = ct + x with respect to a given inertial system.
24
A. Particle history and proper velocity.
1 dX dt 1 dx 0
.
V= = + ,
c dτ dτ c dτ
dx
v= in a given inertial system.
dt
v dt v2
Exercise: Prove V =γ 1+ , γ= = 1− 2
c dτ c
25
For a given observer (free particle) the position x
and time t of a given event X can (in principle) be
determined by radar ranging with light signals, as
illustrated in the figure. Note that the event X is
simultaneous with event X 2 = ct . This geometric
construction is Einstein’s operational procedure for
synchronizing clocks at distant events.
2t = V τ + V˜τ
Exercises:
(1) Compare ages of the twins when the trip is over. Discuss implications of this result.
(2) Prove that the longest path between two points separated by a timelike interval is a
straight line. (See next figure)
26
B. Lorentz contraction: A rod of rest length l0 moves with velocity v who measures its length
as l at time t, as shown in the spacetime map.
l0 = γ l
C. Doppler shift.
v ω′ 1
A distant source with velocity V = γ 1 + emits light signals with frequency f ′ = =
c 2π ∆t ′
ω 1
that are received with frequency f = = , as shown in the figure.
2π ∆ t
27
Exercise: From the figure derive the equation
λ (X 2 − X1 )= D − V ,
D≡ f′ / f
Two different inertial systems (primed and unprimed) with a common origin X 0 = X 0′ = 0 assign
v
different labels X ′ = c t ′ + x ′ and X = ct + x to each spacetime event. If V = γ 1 + is the
c
proper velocity of the primed frame with respect to the unprimed frame, then, in the timelike
plane containing both time axes, the labels are related by the Lorentz transformation:
(1) X = V X′ or X ′ = V˜ X
v⋅ x
t′ = γ t −
c2
x′ = γ (x − vt)
Velocity composition:
Let X ′ = X ′(τ ) and X = X(τ ) represent the history of a particle with proper velocity
u′ u
U ′ = γu′ 1 + and U = γ u 1+ in the two inertial systems.
c c
Exercise: Derive and interpret the relativistic velocity composition law: U ′ = V˜U .
Therefrom, derive the corresponding composition laws for time dilations and relative velocities:
u⋅ v u−v
γ u ′ = γ uγ v 1− , u′ = u⋅ v
c2 1− 2
c
28
Active Lorentz transformations.
v a
V =γ 1+ = e = cosh a + sinh a
c
t v x
cosh a = cosh a = γ = sinh a = vˆ sinha = γ =
τ c cτ
29
mc 2
Energy: E = mc 2γ = = mc 2 + K
2 2
1− v / c
4
2 2 1 v
Kinetic energy: K = (γ − 1)mc ≈ 2 mv + 8 m 2 +
1 ...
c
E
B. The photon is a massless particle with proper momentum P = + p, where the energy is
c
given by Planck’s Law: E = ω = hf .
C. Energy-momentum conservation
P= Pk = Pk
before after
Examples:
Conservation: P1 + P2 = P3 + P4
E h
Photons: P1P˜1 = 0 = P3P˜3, p= p = = hf =
c λ
E2
Electron initially at rest: P2 = = mc
c
30
E E3
Photons: P1 = 1 (1 + pˆ 1) , P3 = (1 + pˆ 3)
c c
[
P4 P˜4 = [(P1 − P3) + P2 ] ( P˜1 − P˜3) + P˜2 ]
m2c 2 = −2 P1P˜3 + 2 ( P1 − P3) P˜2 + m 2c 2
h 2c 2 hc hc
(1 − cosθ ) = − mc 2
λ1λ 2 λ1 λ2
h
λ 2 − λ1 = ∆λ = (1− cos θ) Compton’s formula
mc
P = P1 + P2
Zeroth Law: Make your own formulation to supercede the Newtonian version!
31