MARCEL RIESZ'S WORK ON

CLIFFORD ALGEBRAS
P. LOUNESTO
In,titute 0/ Mathematic,
Hel,inki Unil1er,ity 0/ Technology
SF-O!150 ESPOO, Finland
Abstract. This article reviews Marcel Riesz's lecture notes on Clifford Number, and
Spinor" 1958, and evaluates its effect on present research on Clifford algebras.
The article begins with a critical survey of the history of Clifford algebras. Inaccuracies
in citations are pointed out and mistaken priorities are rectified. Misconceptions about
Clifford algebras are examined and flaws are corrected. The paper focuses on controversial
issues that have been debated over the past few years, and offers definitive resolutions.
Concrete counter-examples are given to misplaced claims.
Particular attention is directed to the existence of a canonical linear isomorphism
between the exterior algebra and the Clifford algebra. ChevalIey constructed the Clifford algebra Cl(Q) of a quadratic form Q as a subalgebra of the endomorphism algebra
End(A V) of the exterior algebra A V. This construction depends on an arbitrary, not
necessarily symmetric, bilinear form B such that B(x,x) = Q(x). The choice of B
fixes the contraction u J 'II in A V and permits the introduction of a Clifford product
xu = x J u + x "u of x E V and u E A v. This gives rise to the Clifford algebra of the
symmetric bilinear form HB(x,y) + B(y, x when the characteristic =1= 2.
Marcel Riesz went backwards with his fundamental formulas
x J u=

-'li.x) and x" u =

+ 'li.x)

and re-introduced the contraction and the exterior product in the Clifford algebra, and
constructed, in alI characteristics =1= 2, a privileged linear isomorphism from the Clifford
algebra to the exterior algebra (without antisymmetric products of vectors). [In the above
'Ii. means the grade involute of u, that is, for a homogeneous element a EA V we have
i. = (-l)a.]
Key words: Exterior algebra - contraction - Kiihler-Atiyah isomorphisms

A History of Clifford Algebras

0.1. ALGEBRAS OF HAMILTON, GRASSMANN AND CLIFFORD
The first step towards a Clifford algebra was taken by Hamilton in 1843
(first published 1844), when he studied products of sums of squares and
invented his quaternions for multiplicative compositions of triplets in 1R3 .
The present formulation of vector algebra was extracted from the quaternion
product of triplets/vectors xy = -x y + x X y by Gibbs 1881-84 (first
published 1901). Hamilton regarded quaternions as quotients of vectors and
215

216

P. LOUNESTO

wrote a rotation in the form y = ax using a unit quaternion a E 1lI, lal = 1.

However, in such a rotation the axis a = a - Re( a) had to be perpendicular
to the vector x (which turned in the plane orthogonal to a). The formula
for rotations x --+ axa- l around an arbitrary axis a E 1R3 was presented by
Cayley in 1845. [Cayley 1855 also discovered the quaternionic representation
q --+ aqb of rotations in 1R4 equivalent to the decomposition Spin( 4) :::'
Spin( 3) X Spin( 3).) Cayley thus came into contact with half-angles and spin
representation of rotations in 1R3. However, already Olinde Rodrigues 1840
had recognized the importance of half-angles in his study on composition of
rotations in 1R3.
The quaternion algebra 1lI is both the even sub algebra and a proper
ideal of the Clifford algebra CiO,3 on 1R3. Hamilton also considered complex
quaternions d:: 1lI isomorphic to another Clifford algebra Ci3 = Ci 3 ,o on
1R3. Note the algebra isomorphisms CiO,3 :::' 1lI a H and Ci3 :::' M 2 (d::).
Bivectors were introduced by H. Grassmann, when he created his exterior
algebras in 1844. The exterior product of two vectors, the bivector a 1\ b,
was interpreted geometrically in n dimensions as the parallelogram with a
and b as edges, and two such exterior products were equal if their parallelograms lay in parallel planes and had the same area with the same sense
of rotation (from a to b). Thus the exterior product of two vectors was
anticomnmtative a 1\ b = - b 1\ a. Using a basis el, e2, ... , en for 1Rn the
exterior algebra ,,1Rn had a basis
1
el, e2, ... , en
el 1\ e2, el 1\ e3, ... , el " em e2 1\ e3, ... , en-l 1\ en

and was thereby of dimension 2n.

W. K. Clifford studied compounds/products of two quaternion algebras,
where quaternions of one algebra commuted with the quaternions of the
other algebra, and applied exterior algebra to grade these tensor products
of quaternion algebras and thus coined his geometric algebra in 1876
(first published in 1878). Clifford's geometric algebra was generated by n
orthogonal unit vectors ell e2, ... , en which anti commuted eiej = -ejei
(like Grassmann) and satisfied all ef = -1 (like Hamilton) [or then all
el = 1 as in Clifford's paper 1882 - the positive definite case is missing in
Budinich&Trautman 1988 p.2 r.20). The number of independent products
Clifford summed
esej = ei 1\ ej, i < j, of degree 2 was In(n - 1) =
up the numbers of independent products of various degrees 0, 1, 2, ... , n

RIESZ'S CLIFFORD ALGEBRA

217

and thereby determined the dimension of his geometric algebra to be

He distinguished four classes of these geometric algebras characterized by

the signs of (ele2 ... en)ei = ei(ele2 ... en) and (ele2 ... en)2 = 1.
He also introduced two algebras of lower dimension 2n - l : the subalgebra
of even elements and, for odd n, a reduced algebra obtained by putting
ele2. en = 1 (instead ofletting ele2 ... en = el/\ e2 /\ ... /\ en to be of
degree n with the requirement (ele2 .. . e n )2 = +1).
0.2.

OFTEN FORGOTTEN: LIPSCHITZ, VAHLEN AND

E.

CARTAN

Clifford's geometric algebra was reinvented in 1880, just two years after
its first publication, by Rudolf Lipschitz, who later acknowledged Clifford's
prior discovery in his book, see R. Lipschitz: Untersuchungen tiber die Summen von Quadraten, 1886 [Budinich&Trautman 1988 p. 2 r. 18 refer to 1886
but seem to be unaware of 1880]. In his study on sums of squares Lipschitz
considered a representation of rotations by complex numbers and quaternions and generalized this to higher dimensional rotations in 1Rn. Lipschitz
thus gave the first geometrical application of Clifford algebras in 1880 [Budinich&Trautman 1988 p. 54 r.10 again refer only to 1886]. He expressed a
rotation y = (1 +A)(1 -A)-IX (written here in modern notation with an antisymmetric matrix A) in the form y - Ay = x + Ax or as y +y. B = x+ Bx
where BEN 1Rn is the bivector determined by Ax = Bx (= -x B). He
rewrote this expression with an even Clifford number a E Clci,n in the form
ya = ax, thus representing the rotation as y = axa- l . (Lipschitz wrote
yal = az where z = xe l \ y = yell, al = elael l .) In modern terms, he
used the exterior exponential a = 1 + B + B /\ B + . .. of the bivector
B so that the normalized element a/lal was in the spin group Spin(n).
[Delanghe&Sommen&Soucek 1992 p.127 follow here Helmstetter 1986 and
attribute these results incorrectly to Sato&Miwa&Jimbo 1978.)
Vahlen 1897 found an explicit expression for the multiplication rule of
two basis elements in Clo,n
e "'le"'2
I 2

e"'''ef31ef32
n
I 2

B.. - (_1)2:i>j"'i{3je"'1+f31 e "'2+f32

e;;'.
I
2

e",,,+{3,,
n

where the exponents are 0 or 1 (added here modulo 2, though for Vahlen
1+ 1 = 2 so that the summation was over i > j) - a formula reinvented frequentlyafterwards [Brauer&WeyI1935, Deheuvels 1981 p. 294, or disguised
with index sets as in Brackx&Delanghe&Sommen 1982 p. 2, Carmichael 1984
p. 228, Chevalley 1946 p. 62, Artin 1957 p. 186, or then hidden among permutations like in Delanghe&Sommen&Soucek 1992 pp. 58-59 who gratuitously

218

P. LOUNESTO

attribute this result to Kiihler 1960/62]. In 1902 Vahlen initiated the study
of Mobius transformations of vectors in 1Rn (or paravectors in 1R +1Rn) by
2 X 2-matrices with entries in Clo,n. This important study was re-initiated
by Ahlfors in the 1980's.
E. Cartan 1908 p. 464 identified the Clifford algebras Clp,q as matrix
algebras with entries in 1R, <C, 1H, 1RE91R, 1HE91H and found a periodicity of
8 (a re1inement of Clifford's original 4). Cartan also observed spinor modules
of orthogonal Lie algebras in 1913 (though he did not introduce the term
"spinor"). Spinors got their current name after they were reinvented and
applied to the spinning electron by Pauli 1927 and Dirac 1928.
Two-valued spinor representations of rotations in n-dimensions were reconstructed recursively by Brauer&Weyl in 1935. They applied Clifford algebras without using the term "Clifford algebra" [misleading wording in
Chevalley 1954 p.4, Crumeyrolle 1990 p. xi]. Brauer&Weyl also reinvented
(in the case of the Clifford algebra Cln ) the above explicit multiplication
formula of Vahlen, but failed to observe the connection to the Walsh functions (discovered in the meanwhile by Walsh 1923). This connection to the
Walsh functions was brought to general attention by Hagmark&Lounesto
1986.
0.3.

MAXWELL EQUATIONS AND DIRAC SPINORS

Maxwell's equations were condensed into a single equation by complex vectors (Silberstein 1907), complex quaternions (Silberstein 1912/1914), spinors
(Laporte&Uhlenbeck 1931) and by Clifford algebras (Juvet&Schidlof 1932,
A. Mercier 1935) - [mistaken priority in Hestenes 1966 p. 29]. In the Clifford
algebra Cia,}
M4(1R) the Maxwell equations were written as {)F = J
where {) = V - eo{)o and F = Eeo - Be12a E A21Ra,l. The single Maxwell
equation could be decomposed into two parts, {) A F = 0 and {). F = J. Similarly, {)A = -F could be decomposed into two parts, {)AA = -F and the
Lorenz condition {). A = 0 [discovered along with the Lorenz gauge by the
Danish physicist Ludwig Lorenz and not the Dutch physicist H. A. Lorentz,
who invented the Lorentz metric and the Lorentz transformations, see J. van
Bladel: Lorenz or Lorentz? IEEE Antennas and Propagation Magazine 33
(1991) p.69 and The Radioscientist 2 (1991) p.55]. Marcel lliesz 1947 expressed the Maxwell stress tensor as T/J" = - t(e/JFe"F)o and Hestenes 1966
p. 31 introduced the vectors T/J = -tFe/JF for which T/J" = T/Je" = e/J.T"
and the mapping Tx = -tFxF where (Tx)/J = T/J"z" [Juvet&Schidlof
1932 p. 141 gave
"'F.
1
F. F",{J
T/J&J = F/J "''' + 4" g/J" "'/3

=-

but did not consider Tx

t FxF - this revises a citation in Bolinder 1986
p. 469]. Note that To = HE2 + B2)eo + E X B.

219

RJESZ'S CLIFFORD ALGEBRA

Juvet 1930 and Sauter 1930 replaced column spinors by square matrices
in which only the first column was non-zero - thus spinor spaces become
minimal left ideals in a matrix algebra. Riesz 1947 used primitive idempotents of Clifford algebras to construct spinor spaces as minimal left ideals
in Clifford algebras [Budinich&Trautman 1988 p. 12 misplaced priority to
Chevalley 1954]. Hestenes 1966 formulated the Dirac theory with the real
Clifford algebra Gl 1 ,3 M2(H) of the Minkowski time-space lR1 ,3 - this
led him in 1966/67 to the lucid but at first sight perplexing idea of replacing spinors spaces by the even subalgebra elt 3 of the real Clifford algebra
Gl1 ,3.
'

0.4.

THE ROLE OF

E.

WITT

E. Witt 1937 used Clifford algebras in his study on quadratic forms over arbitrary fields of characteristic :j:. 2. Witt's paper began the modern algebraic
theory of quadratic forms, and recognized Clifford algebras as algebras of
quadratic forms. The Witt ring W(K), of a field K, consists of similarity classes of non-singular quadratic forms over K (similar quadratic forms
have isometric anisotropic parts). In characteristic :j:. 2 the structure of Clifford algebras of some special quadratic forms was studied by Lee 1945/48
(el = 1), by Chevalley 1946 (el = -1), and by Kawada&Iwahori 1950
(el = 1) without the benefit of the Witt ring (and so they did not consider all the isometry classes of anisotropic quadratic forms).

Example. The Witt ring W{Fs) of the finite field Fs

{O, 1, 2, 3,4} of
characteristic 5 contains four isometry classes 0, (1), (s), (1, s) where s = 2
or s = 3. Chevalley 1946, Lee and Kawada&Iwahori had not noticed that
none of the quadratic forms
+
on the plane Fl
is isometric with
+
(1, s) (but in fact all are neutral, and thereby
in the same isometry class as 0). A simpler counter-example is the line Fs
where the Clifford algebra of 2;:c2
(2) is the quadratic extension Fs( V2)
whereas the Clifford algebras of ;:c2
(1) split Fs X Fs.

0.5.

ISOMORPHISMS OF CHEVALLEY,

M.

RIESZ AND KAHLER-ATIYAH

Several important generalizations were introduced by C. Chevalley, 1954.

He emphasized that Clifford algebras are algebras of quadratic forms (but
did not acknowledge Witt), allowed arbitrary quadratic forms (as Witt)
and ground fields even of characteristic 2. Chevalley introduced the exterior
exponential of bivectors and used it to scrutinize properties of Lipschitz's
covering group of rotations (naming it unjustly a 'Clifford group'). In case
of a neutral quadratic form Chevalley induced spinor spaces by volume elements of a totally null subspace, and studied Cartan's pure spinors and
triality (in dimension 8). More importantly, Chevalley identified Clifford al-

220

P. LOUNESTO

gebras as sub algebras of the endomorphism algebra of the exterior algebra

(he needed this to allow the ground field to be of characteristic 2).
In his monograph entitled Clifford Numbers and Spinors, 1958, pp. 61-67,
M. lliesz introduced a second product into Clifford algebras making them
isomorphic with the exterior algebra. This construction is valid in all characteristics f:. 2 (as opposed to Chevalley's 1946 construction by completely
antisymmetric products of vectors valid only in characteristic 0). [In spite
of its title lliesz's monograph did not deal with spinors - misleading information in Brackx&Delanghe&Sommen 1982 p. 40, Carmichael 1984 p. 229.]
In 1964 Atiyah&Bott&Shapiro related the spin groups to K-theory when
they re-identified definite real Clifford algebras Cin and Cin,o as matrix
algebras with entries in JR,
H, JR E9 JR, H E9 H. They rediscovered the
periodicity of 8 (found by Cartan in 1908) with respect to the graded tensor
product (used earlier by Chevalley). They emphasized the importance of the
Z2-graded structure (the even-odd parity in Clifford's works) and used it to
simplify Chevalley's approach to Lipschitz groups (the role of parity grading
and/or grade involution in reflections was not observed by Chevalley). They
also wanted to reconsider spinor spaces as modules over a Clifford algebra
(instead of regarding them as minimal left ideals in the Clifford algebra).
This paper seems to be best known for its dirty joke: the pronunciation
of 'pin group' in English sounds the same as 'pine groupe' in French (your
audience will laugh if you talk about 'pine groupe' in France). Hestenes went
on with this joke by observing that a 'pin group' consists of two parts - the
even = spin and the odd part.
E. K8.hler 1960/62 introduced a second product for Cart an's exterior
differential forms making Grassmann's exterior algebra isomorphic with a
Clifford algebra. This re-interpretation of Chevalley's Clifford algebra was
renamed Kiihler-Atiyah algebra by W. Graf 1978 (Graf's K8.hler-Atiyah algebra was again reinvented and applied to the K8.hler-Dirac equation by
Salingaros&Wene 1985). However, lliesz's idea of going backward and introducing exterior product in Clifford algebra is simpler than Chevalley's
Clifford algebra or Graf's Kahler-Atiyah isomorphism.
Many researchers have taken it for granted that bivectors and dimension grading exist in Clifford algebras. However, this requires the existence
of a canonical linear isomorphism between Clifford algebras and the exterior algebra, but the existence of such a privileged isomorphism has been a
controversial issue to which I hope to give a definitive answer in the next
section. (It is generally agreed that Clifford algebras admit a parity grading with even and odd parts Cl(Q) = Cl+(Q) + Cr(Q), but opinions are
divided over the existence of a dimension grading with k-vector parts
Ci( Q) = K + V + N V + ... + V. We shall see that there is a natural
choice for the dimension grading in all characteristics f:. 2.)

"n

221

RIESZ'S CLIFFORD ALGEBRA

Marcel Riesz's 'Clifford Numbers and Spinors'

1. Chapter 1: 'Clifford Numbers'

Consider an n-dimensional vector space V over a field K, char K =F 2, and

the symmetric bilinear form
<x,y> =

'2 (Q(x + y) -

Q(x) - Q(y))

associated with the quadratic form Q on V. Occasionally I shall discuss

the case char K = 2. I will denote the exterior algebra of V by A V and
the Clifford algebra Cl of Q, x 2 = Q(x), by Cl(Q) or by Clp,q in case of
V = 1RP ,q with a non-degenerate quadratic form
Q(x)

+ ... + z; - z;+1 - ... - z;+q,

=p + q.

I shall construct a natural linear isomorphism A V

-+ Cl(Q), review how

Riesz goes backwards using an identification Cl( Q) -+ A V, and compare
Riesz's method to an alternative construction due to Chevalley but known to
some theoretical physicists in the disguise of the Kahler-Atiyah isomorphism.

1.1. CONSTRUCTION OF THE LINEAR ISOMORPHISM

AV

-+

Cl(Q)

Start from the exterior algebra A V. Define for a k-vector a E AIe V the
grade involution a = (-l)lea and the reversion ii = (_l)tle(le-l)a. The
symmetric bilinear form associated with Q on V can be extended to simple
k-vectors in Ale V by way of <Xl A X2 A A XIe,Yl A Y2 A . A n> =
det<xi,Yj>, where
<Xt,Yl>
<X2,Yl>

<Xt,Y2>
<X2,Y2>

and further by linearity to all of AIe V and by orthogonality to all of A V.

Example. Let Q(zlel +Z2e2) =
Then <x,y> = aZlYl +bZ2Y2
and x A y = (ZIY2 - z2ydel A e2' The identity
+
+ byn =
(azlYl +b:z:2Y2)2+ab(ZIY2-Z2Yl)2 can be written as Q(x)Q(y) = <x, y>2+

Q(x A y), where Q(x A y) = ab(zlY2 - z2yd 2.

In case of a non-degenerate Q on V we can introduce the dual of the exterior

product called the contraction u J v of v E A V by u E A V with the
requirement

<u J V,w> = <v,u A w>

for all wE /\ V.

P. LOUNESTO

222

Examples. 1. Let X,y E V, W E K. Then <x J y,w> = <y,x /\ W> =

<y,xw> = <y,x>w and since <x J y,w> = <x J y,l>w we have the
rule x J y = <x,y>.

2. Let x,y,z,WEV. Then <xJ(y/\z),w>=<y/\z,x/\w>

=

I <y,x>
<z,x>

<y,W>
<z,w>

= (x J y)<z, W> - (x J z)<y, W> = x J y)z - (x J z)y, w> and so we

have the rule x J (y /\ z) = (x J y)z - (x J z)y.

3. Let x E V, Xi E V and w = Wl /\ W2 /\ ... /\ Wk-l E Ak-l V. Then

<x J (Xl/\X2/\ .. . /\Xk),W> = <Xl/\X2/\.' ./\Xk,X/\ Wl/\ W2/\"'/\ wk-l>
<Xl, x>
<X2,X>

<Xl, Wl>
<X2,Wl>

<Xl, W2>
<X2,W2>

<Xl, Wk-l>
<X2, Wk-l>

'L) _1)i-l<x, Xi><Xl/\ X2/\"'/\ Xi-l /\ Xi+!/\ ... /\ Xk, w>

i=l
and so

-1)i-l<X,Xi>Xl/\ X2 /\ ... /\ Xi-l/\ Xi+!/\"'/\ Xk

i=l
4. x J (u /\ v) = X J ((Ul/\ U2 /\ ... 1\ Ui) 1\ (Vl 1\ V2 /\ ... 1\ Vj))
i

Uk)( -1)k-l ul

/\ U2 1\ 1\ uk-l 1\ Uk+! 1\ 1\ Ui /\ v

k=l
j

+( _1)i

J Vk)( _l)k-l u 1\ Vl 1\ v2 1\ /\ Vk-l /\ Vk+l 1\ 1\ Vj

k=l

= (x J u) 1\ v + (-I)iu 1\ (X J v).

vr /\

5. u/\v) J w,z> = <w,(u/\

z> = <w,ii/\ u/\ z> = <v J w,u/\ z>
= <u J (v J w), z> and so (u /\ v) J w = u J (v J w).

In case of a non-degenerate Q we have verified the following properties of

the contraction

223

RIESZ'S CLIFFORD ALGEBRA

(a)

x J y = <x, y> for x, y E V

(b)

xJ(uAv)=(xJu)Av+uA(xJv)

( c)

( U A v) J w = u J (v J w)

for u, v, w E AV

see Hehnstetter 1982. These properties uniquely detennine the contraction

also for an arbitrary, not necessarily non-degenerate Q. The identity (c)
introduces a scalar multiplication on AV making it a left module over AV.
The identity (b) means that x E V operates like a derivation. Evidently,
x J a E AIe-I V for a E AIe V.
Introduce the (Clifford) product of x E V and u E AV by
XU= xAu+x J u
and extend this product by linearity and associativity to all of AV making it
isomorphic to Cl( Q). For instance, the product of a simple bivector x AyE
A2 V and an arbitrary element u E AV is given by
(x A y)u = x A Y A u + x A (y J u) - Y A (x J u) + x J (y J u)
where we have first expanded (x 1\ y)u = (xy - x J y)u then used
x(yu) = x A Y A u + x 1\ (y J u) + x J (y AU) + x J (y J u)
and the derivation rule x J (y A u) = (x J y) Au - Y A (x J u).
Exercises. Show that
1. XA(yJu)-YA(xJu)=xJ(YAu)-yJ(xAu) for x,yEV
2. x A Y A (z J u) - x A z A (y J u) + y A z A (x J u)
= x J (y A z A u) - y J (x A z A u) + z J (x A Y AU)
3. x A (y J (z J u)) - y A (x J (z J u)) + z A (x J (y J u))
= (x A y) J (z Au) - (x A z) J (y A u) + (y A z) J (x AU)
= x J (y J (z 1\ u)) - x J (z J (y 1\ u)) + y J (z J (x A u))
4. (x A Y A z)u = x A Y 1\ (z J u) - x A z A (y J u) + Y 1\ z 1\ (x J u)
+x A (y J (z J u)) - y A (x J (z J u)) + z A (x J (y J u))
+x A y A z Au + x J (y J (z J u))
5. a J b E Ai - i V for a E Ai V, b E Ai V
In the last exercises we have a non-degenerate Q.
Define u L v by <u L v,w> = <u,w A ii>
(we say that v contracts u and also u is contracted by the contractor v).
6. u J v = Va L Ua - Va LUI + VI L Ua + VI LUI = V L u - 2va LUI
(Va = even(v), UI = odd(u))

224
7.

P. LO UNESTO
u

Lv =

Vo

J tto

+ Vo J

ttl - VI

(VI = odd(v), Uo = even(u))

8. a J b = (_l),U-i)b L a for a E

Uo

,,'V,

9. a E ,,10 V, x E V, x J a

=0

10. b E /\,. V is simple

(a J b) A b

:>

+ 111 J

UI

11

J u - 2V1 J

b E I\ i V

x = a J b for some b E

=0

Uo

,,"'+1 V

for all a E "Ic-l V.

Remark. The dot-notation a b may be used for the contraction, when

it is clear from the context which one of the factors is contracted and which
is the contractor. This dot product a' b can be used, when at least ODe
of the factors is homogeneous. If both the factors are homogeneous, then we
agree that the one with lower (or not higher) degree is the contractor
.:

ab = a J b for a el\v,

be/\V, i:5:i

.:

ab = aLb for a e /\v, b E !\V,

and when precisely one fact or i s known to be homogeneous we agree that it
is the contractor

au = aJu for a e /\v, u E I\V

;

u-a=uLa for a E I\V, U EI\V_

Note that the contraction obeys the rules 1 J u tI, tt E 1\ V, and x J 1 0,

x E V, whereas the dot product obeys the rule 1 - tI = U - 1 = u . Note also
that the dot product does not act like a scalar multiplication on the left
1\ V-module 1\ V, that is, (a li b) u i:- a (b ttl. As an exercise the reader
may verify that for a = el, b = el + ez and u = e l + el II e2 in 1\ IR.2 all
the expressions

a J (b Ju), a J (b L u), a L(b J u), a L(b Lu) and

(aA b) Ju, (a A b) Lu
are unequal with the exception of (a li b ) J u= a J (b J u) = ?
- The lack of linearity over /\ V renders less useful any extension of u v
for arbitrary u, v E /\ V. Such an extension was actually introduced under the name of 'hmer product' by Hestenes&Sobczyk 1984 p.6 who also
demonstrated its futility by displAying only formulas with at least one homogeneous factor. The non-/\ V -linear 'iruler product ' is not consistent with
the contraction (in the sense that the 'inner product' is not a special case
of the contraction), because tt v might differ simultaneously both from
u J v and u Lv. Boudet 1992 p. 345 mentioned a formalization of the

225

RIESZ'S CLIFFORD ALGEBRA

'inner product' but his rules are not sufficient to permit the evaluation of
(X 1\ y) . u when U E /\ V, U rf. V (though they do permit a construction of
the 'inner product' with an additional rule (x 1\ y) . U = X (y . u) where
u E /\k V, k ;::: 2).

The above remark shows how the asymmetric contration solves a problem
of Hestenes&Sobczyk 1984, who postulate the 'inner product' to be 0 [p.6,
r. 12 formula (1.21b)) if one of the factors is a scalar, and run into difficulties
on p. 20 rows 8-18 formula (2.9). However, as the following example shows
the problem is deeper than that since the 'inner product' is not equal to the
contraction even though scalars would be excluded.
Example. Let el, e2 be an orthonormal basis for R 2. Compute

in the sense of Hestenes&Sobczyk. The same elements have the contractions

(el - el 1\ e2) J (e2
(el - el 1\ e2) L (e2

+ el 1\ e2) = 1 + e2

+ el 1\ e2) =

1 - el.

This shows that neither of the contractions coincide with the 'inner product'
of Hestenes&Sobczyk.

Remark. We may re-obtain the dot product in terms of the Clifford product
as follows (char i- 2)
a b

= (ab)li-il

for a E /\ V, bE /\ V

where (U)k is the k-vector part of U E /\ V ::: Cl( Q). For two homogeneous
elements a E /\i V, b E /\i V the Clifford product decomposes into k-vector
parts
ab = (ab)i+i + (ab)i+i-2 + ... + (ab)li-il
where al\b = (ab)i+i and ab = (ab)li-il' In particular, ab =I al\b+aJb,
when i =11 even for i < j, as can be seen by the counter-example
a

2 3 3
ab = e134 E /\R 4

= e12 E /\R\ b = e234 E /\R\

[this blunder lurks in Salamon, 1989, p. 170).

1.2.

CHEVALLEY'S IDENTIFICATION OF

Cl(Q) C End(/\ V)

Chevalley, 1954, pp. 38-42, introduced a linear operator

such that
<Px ( u) = x 1\ U + x J u for x E V, u E /\

= <px E End(/\ V)

v.

226

P. LOUNESTO

From the derivation rule x J (x /\ u) = (x J x) /\ u - x /\ (x J u) and

x/\x/\u = 0, xJ(xJu) = 0 Chevalley concluded the identity (<Px)2 = Q(x).
His inclusion map V -> End(A V), x -> <Px was then a Clifford map and
could be extended to an algebra homomorphism 1/J: Cl(Q) -> End(A V),
whose image evaluated at 1 E A V yielded the map : End(A V) -> A V.
The composite linear map () = 0 1/J was the right inverse of the natural
map A V -> Cl(Q) and

/\ V

->

Cl(Q)

t End(/\ V) t /\ V

was the identity mapping on A V. The faithful representation 1/J sent Cl( Q)
onto an isomorphic subalgebra of End(A V).
Chevalley's identification works fine also with a contraction defined by
an arbitrary - not necessarily symmetric - bilinear form B such that
B(x, x) = Q(x) and

= B (x, y)

for x, y E V

( a)

x J y

(b)

xJ(u/\v)=(xJu)/\v+u/\(xJv)

(c)

(u /\ v) J w

=u J

(v J w) for u, v, w E A V

see Helmstetter 1982. As before x J a E Ak- l V for a E Ak V and

x J (Xl /\ x2 /\ ... /\ Xk)
k

= Z)-1)i- 1 B(x,Xi)Xl/\ X2/\ ... /\Xi-l/\Xi+1/\ ... /\ xk

i=l
and the faithful representation 1/J sends the Clifford algebra Cl( Q) onto an
isomorphic subalgebra of End(A V), which, however, as a subspace depends

onB.
Remark. Chevalley introduced his identification Cl(Q) C End(A V) in
order to be able to include the exceptional case of characteristic 2. In characteristic f= 2 the theory of quadratic forms is the same as the theory of
symmetric bilinear forms and Chevalley's identification gives the Clifford
algebra of the symmetric bilinear form <x,y> = t(B(x,y) + B(y,x))
satisfying xy + yx = 2<x, y>.
I
For arbitrary Q but char K f= 2 there is the natural choice of the unique
symmetric bilinear form B such that B(x,x) = Q(x) giving rise to the
canonical/privileged linear isomorphism Cl( Q) -7 A V. The case char K =
2 is quite different. In general there are no symmetric bilinear forms such
that B( x, x) = Q (x) and in the case that there is such a symmetric bilinear
form, then it is not unique, since any alternating bilinear form (is also symmetric and) could be added to the symmetric bilinear form without changing

227

RIESZ'S CLIFFORD ALGEBRA

Q. [Recall that antisymmetric means B(x,y) = -B(y,x) and alternating

B(x, x) = OJ alternating is always antisymmetric, though in characteristic
2 antisymmetric is not necessarily alternating.] Thereby the contraction is
not unique, and there is an ambiguity in I{)x.
In characteristic 2 the theory of quadratic forms is not the same as the
theory of symmetric bilinear forms.
Example. Let dimV = 2, B(x,y) = aZlYl + bz l Y2 + CZ2Yl + dZ 2Y2 and
Q(x) = B(x,x). The contraction xJy = B(x,y) gives the Clifford product
xv = x A v + x J v of x E V, v E "V. We will determine the matrix of
v - uv, U = Uo + Ul el + U2e2 + U12el A e2 with respect to the basis
1, ell e2, el A e2 for
The matrix of el is obtained by the following
computations

"V.

= 0100)
= aOOO)

(first column

= el J el = a
(second column
ele2 = el A e2 + el J e2 = el A e2 + b
(third column = bOOl)
el (el A e2) = el J (el A e2) = (el J ed A e2 - (el J e2) A el = ae2 el el

bel

and the matrix of el A e2 by

(el A e2) 1 = el A e2

= (el A e2) L el = el A (e2 L ed - e2 A (el L el) = eel (el A e2)e2 = (ele2 - el J e2)e2 =
- (el J e2)e2 = del - be2
(el A e2)(el A e2) = (ele2 - el J e2)(el A e2)
= el(e2 A (el A e2) + e2 J (el A e2 - (el J e2)(el A e2)
= el(ee2 - ded - b(el A e2) = -ad + be + (-b + c){el A e2)'
(el A e2 )el

So we have the following matrix representations

b
0 0
0 0
0 1

e
0

e,

a
0

0
el A ea

e
-a

0
or in general
U

C
Ul
U2
U12

'

0
d
-b
0

aUl + eU2
Uo + eU12
- au 12
-U2

e,

0
0 -1

d
0
0
0

-aYbO)
-b+ e
bUl + dU2
dU12
Uo - bU12
Ul

-(ad
- bolu"
-(bUl + dual
Uo

aUl + cU2
+ (-b + e)ul2

ae2

228

P. LOUNESTO

Evidently, the conunutationrelations ele2+e2el = b+e and ei

are satisfied, and we have the following multiplication table

:f

=d

-bel + ae2
-del + ee2
-ad + be + (-b + e)el 1\ e2

-el 1\ e2 + e
eel - ae2

In characteristic

= a,

2 we find
1
1
'2 (el e2 - e2 e d = el 1\ e2 + '2 (b - e)

and more generally for x = xlel + X2e2, y = Ylel + Y2e2

1

'2 (xy -

yx) = (XlY2 - x2yd el 1\ e2 +

'2 (b -

e)(xlY2 - x2yd

= x 1\ Y + A( x, y)

with an alternating scalar valued form A(x,y) = t(B(x,y) - B(y,x)). For

non-zero A(x,y) the quotient x 1\ y/A(x,y) is independent of X,y E V.
Note that the matrix of Hxy - yx) is traceless. The synunetric bilinear
form associated with Q(x) is
X Y =

'2 (B(x,y) + B(y,x)) =

aXlYl +

'2 (b + e)(xlY2 + x2yd + dX2Y2

and we have xy + yx = 2x . y for X,y EVe Cf(Q). Orthogonal

vectors x..l y anticonunute xy = -yx and (xy)2 = _x 2y2 even though
xy = x 1\ Y + A(x,y). (In this special case A(x,y) = B(x,y) :f 0 while
xy = 0 implies B(y,x) = -B(x,y).)

It is convenient to regard A V as the subalgebra of End(A V) with the

canonical choice of the synunetric B = o. We may also regard Cf(Q) as a
sub algebra of End(A V) obtained with some B such that B(x, x) = Q(x)
and choose the synunetric B in char :f 2.
The following example shows that for Q = 0 and B = 0, Chevalley's
process results in the original multiplication of the exterior algebra A V but
that for Q = 0 and alternating B :f 0, the process gives an isomorphic but
different exterior multiplication on A V.

Example. Take a special case of the previous example, the Clifford algebra
with Q 0 and B(x,y) b(XlY2 - 2:2Yl). Send the matrix of the exterior
product (with the synunetric bilinear form = 0) to a matrix of the isomorphic
Clifford product (determined by the alternating bilinear form = B)

Uo

229

RIESZ'S CLIFFORD ALGEBRA

In this case
J3(x)J3(y) = J3(x A Y + B(x, y)).

In particular, J3(el)J3(e2) = J3(elAe2+b), J3(e2)J3(ed = -J3(elAe2+b) and

,8( el Ae2 + b),8( e;) = 0, ,8( ei),8( el Ae2 + b) = O. We will meet this situation
again in the section on the 'Uniqueness and the definition by generators
and relations' except that here the exterior algebra and the Clifford algebra
(determined by the alternating B) are regarded as different subspaces but
isomorphic subalgebras of End(A V).

The above example shows that those who refute the existence of k-vectors
in a Clifford algebra Cf(Q) over K, char K f- 2, should also exclude fixed
subspaces A k V c A

v.

In general, consider two copies of Cf(Q) in End(A V) so that Q(x) =

B1(x, x) = B2(X, x) determines J31(X),81(Y) = ,81(X A y + B1(x, y)) and
,82(X),82(Y) = ,82(X A y + B 2 (x,y)). A transition between the two copies is
given by an alternating bilinear form B(x,y) = B1(x,y) - B 2 (x,y) and
,8(x),8(y) = ,8(xy + B(x,y)).
In char K f- 2 this means that the symmetric bilinear form such that
<x,x> = Q(x) gives rise to the natural choice xy = x A Y + <x, y>
among the Clifford products xy + B( x, y) with an alternating B. In other
words, the Clifford product xy has a distinguised decomposition into a sum
xAy+<x,y> ofascalar <x,y>= !(xy+yx) andabivector xAy=
!(xy - yx) among the possible decompositions with antisymmetric part
XAY = !(xy-yx) equaling a new kind ofbivector xAy = xAy+ B(x, y).
[Similarly, a completely antisymmetric product of three vectors equals a new
kind of 3-vector x A y A z = x 1\ Y 1\ z + xB(y,z) + yB(z,x) + zB(x,y).)
Example. Let K = {O,1}, dim V = 2 and Q(xlel + X2e2) = XIX2.
There are only two bilinear forms Bi such that Bi(X, x) = Q(x), namely
B1(x,y) = XIY2 and B 2 (x,y) = X2Yt, and neither is symmetric. The difference A = B 1 -B 2 , A(x,y) = XIY2-X2Yl (= XIY2+X2yd is alternating (and
thereby symmetric). There are two representations of Cf (Q) in End (A V),
namely
for B 1 :

Ul

Ul

Uo

U2

U12

-U2

C'

Uo -

U2

for B 2 :

Ul
U2
U12

Uo

+ Ul2
0
-U2

-"1 )

U12

Ul

Uo

0
0

o0 )

Uo
Ul

U12

U2
Uo

U12

These representations have the following multiplication tables

230

P. LOUNESTO

Bl
el
e2
el A e2

B2
el
e2
el A e2

el A e2
el
e2
0
-el
1 + el A e2
0
0
-el A e2
0
-e2
-el A e2
e2
el A e2
el
0
0
el A e2
0
1- el A e2
e2
0
el
el A e2

with respect to the basis 1, el, e2, el Ae2 for " V. In this case there are only
two linear isomorphisms "V -+ Cl(Q) which are identity mappings when
restricted to K + V and which preserve parity (send even elements to even
elements and odd to odd). It is easy to verify that the above multiplication
tables of Cl( Q) are actually the only representations in
In this case
there are no canonical linear isomorphisms
-+ Cl(Q), in other
words, neither of the above multiplication tables can be preferred over the
other. In particular, N V cannot be canonically embedded in Cl(Q), and
there are no bivectors in characteristic 2!

"V "V.

The above example shows that Lawson&Michelsohn pp. 10-11 are wrong
when they allow characteristic 2 and claim that there is a canonical linear
isomorphism" V -+ Cl(Q). Chevalley used unnecessarily complicated and
abstract methods, since he wanted to include the case char K = 2 and
employed endomorphisms of
to get a faithful representation of Cl(Q).
The need for a simplification of Chevalley's presentation is obvious. For
instance, B. L. van der Waerden: On Clifford algebras, 1966, said that 'the
ideas underlying Chevalley's proof (p. 40) are not easy to discern' and gave
another proof, equivalent but easier to follow. (Also Crumeyrolle 1990 p. xi
claims that 'Chevalley's book proved too abstract for most physicists' and
in a Bull. AMS review Lam 1989 p. 122 admits that 'Chevalley's book on
spinors is ... not the easiest book to read. ') It might be helpful to get acquainted with a simpler and more direct method of relating " V and Cl( Q)
due to Marcel lliesz: Clifford Numbers and Spinors, 1958, pp. 61-67. lliesz introduced a second product in Cl( Q) making it isomorphic with" V without
resorting to the usual completely antisymmetric Clifford product of vectors
and constructed a privileged linear isomorphism Cl( Q) -+ " V.

"V

1.3.

RIESZ'S INTRODUCTION OF AN EXTERIOR PRODUCT IN

Cl(Q)

Start from the Clifford algebra Cl( Q) over K, char K -::J. 2. The isometry
x -+ -x of V is extended to an automorphism of Cl( Q) called the grade
involution u -+ U. Define the exterior product of x E V and u E Cl( Q)

231

RIESZ'S CLIFFORD ALGEBRA

by
(xu + ux),

x Au =

u Ax =

(ux + xu)

and extend this exterior product by linearity to all of Cl( Q) which then
becomes isomorphic to AV. The exterior products of two vectors x A y =
(xy - yx) are simple bivectors and they span A2 V. The exterior product
of a vector and a bivector x A B = (xB +Bx) is a 3-vector in A3 V. The
subspace of k-vectors is constructed recursively by

We may deduce associativity of the exterior product as follows. First, the

definition implies for x, y, z E V
1
x A (y A z) = 4" (xyz - xzy + yzx - zyx)

1
(x A y) A z = 4" (xyz - yxz

+ zxy -

zyx).

Then xA(yAz)=(xAy)Az since

xyz - zyx

= xyz -

zyx + (zy

+ yz)x -

x(yz + zy)

= yzx -

xzy.

This last result implies

1
x A Y A z = "6 (xyz + yzx + zxy - zyx - xzy - yxz)

when char K -=f 2, 3 (note the resemblance with anti symmetric tensors).
[Similarly, we may conclude that xyz + zyx = x(yz + zy) - (xz + zx)y +
z(xy + yx) is a vector in V.] lliesz's construction shows that bivectors
exist in all characteristics -=f 2.
Introduce the contraction of u E Cl( Q) by x E V so that
xJ u

(xu - ux)

and show that this contraction is a derivation of Cl( Q) while

x J (uv)

(xuv - Uvx)

= !2 (xuv -

(xuv - uvx)

uxv + uxv - uvx)

= (x J u)v + u(x J v).

232

P. LOUNESTO

Thus one and the same contraction is indeed a derivation for both the exterior product and the Clifford product. (Kahler 1962 p. 435 (4.4) and p. 456
(10.3) was aware of the equations x J (u 1\ v) = (x J u) 1\ v + it 1\ (x J v)
and x J (uv) = (x J u)v +it( x J v).) Provided with the scalar multiplication
( u 1\ v) J w = u J (v J w) the exterior algebra A V and the Clifford algebra
Cl(Q) are linearly isomorphic as left A V-modules.
Exercises. Show that (char K

f:.

2)

1. x and x 1\ y anti commute for vectors x, y E V

2. x and x J B anticommute for a bivector B E A 2 V
3. (x 1\ y)2 = (x J y)2 - x 2y2 (Lagrange's identity)
4. (x 1\ Y 1\ z) J u = (x 1\ y) J (z J u) = x J (y J (z J u

5. (xyz - zyx)2 E K, x 1\ Y 1\ z

= Hxyz -

zyx)

= (_l)iib 1\ a for a E Ai V and bE Ai V

u 1\ v = Vo 1\ Uo + Vo 1\ UI + VI 1\ Uo - VI 1\ UI = v 1\ u -

6. a 1\ b
7.

u,v E A V,

Uo

= even(u),

Vo

= even(v),

UI

= odd(u),

VI

2VI

1\ UI where

= odd(v)

8. Bu = B 1\ u + HBu - uB) + B J u for BEN V

[Hint: (x 1\ y) 1\ u = (x 1\ y) J u = x 1\ (y 1\ u) + x J (y J u)]
9. Q(u) = (uu)o, <u,v> =

(u J v)o (=

the scalar part of

In the last exercise we have a non-degenerate Q:

10. Q on V extends to a neutral or anisotropic Q on Ct.

u J v)

Riesz's construction shows that Crumeyrolle 1990 pp. 42-45, p. 292 is mistaken when he excludes characteristic 2 and claims that there is no canonical
linear isomorphism A V -+ Cl(Q). Crumeyrolle seems to take it both ways
when he agrees that in char f:. 2 there is a natural choice for Cl(Q) c
End(A V) given by the unique symmetric B such that B(x,x) = Q(x).
Crumeyrolle and Lawson&Michelsohn have opposite opinions about the existence of a canonical linear isomorphism A V -+ Cl(Q) but both are wrong:
Lawson&Michelsohn because they allow char 2 and Crumeyrolle because
he excludes char 2. Curiously, Crumeyrolle excludes char 2 even though he
uses Chevalley's method developed specifically for char 2. Neither Crumeyrolle nor Lawson&Michelsohn have understood that Chevalley presented his
regular-type representations of Cl(Q) in End(A V) in order to be able to
handle also the exceptional case of char 2. [This motive of Chevalley has
also escaped Harvey 1990 p. 180 and Budinich&Trautman 1988 p. 58.]
(Roughly speaking, Lawson&Michelsohn claim that two numbers x, y in
a field K always have a mean !(x + y) within that field K, but this is false
in the field of two elements K = {O, I} and Crumeyrolle claims that no

233

RIESZ'S CLIFFORD ALGEBRA

such mean exist in K, but he is wrong while he excludes char 2.)

1.4.

UNIQUENESS AND THE DEFINITION BY GENERATORS AND RELATIONS

The following definition is favoured by physicists:

Definition. An associative algebra over K with unit 1 is the Clifford algebra Cl of a non-degenerate Q on V if it contains V and K = K . 1 as
distinct subspaces so that

= Q(x)

for any x E V

(1)

x2

(2)

V generates Cl as an algebra over K

(3)

Cl is not generated by any proper subspace of V.

The universal property is guaranteed by including the third property (3).

Using an orthonormal basis ell e2, ... , en of R n , generating Clp,q, the
condition (1) can be expressed as

while the condition (3) becomes Porteous' 1969 requirement ele2'" en f:.
1 which is needed only in the signatures p - q = 1 mod 4. The relations
(La) also generate a lower-dimensional non-universal algebra in any signature p - q = 1 mod 4 in which all the basis elements ei commute with
el e2 ... en and (el e2 ... e n )2 = 1. However, no similar counter-examples
exist in even dimensions. Therefore, whereas it is correct to introduce the
Clifford algebra of the Minkowski space-time without the condition (3), in
arbitrary dimensions it is controversial to omit condition (3).
The above definition gives a unique algebra only for non-degenerate (nonsingular) quadratic forms Q. In particular, it is not good for a degenerate
Q as is shown by the following two counter-examples in case Q = O.
1. Define for x, y E V the product xy
K Efl V an algebra of dimension n + 1.

= O.

This makes the direct sum

2. Introduce a product in AR 3 by eiej = ei A ej for all i,j = 1,2,3 and

el e2e3 = O. Thus the subspace R R 3 A2 R 3 of A R 3 is a 7-dimensional
algebra generated by Rand R3.

+ +

This shows that it is not possible to replace the condition (3) by the requirement that only parallel vectors commute. We could include arbitrary
quatratic forms Q by requiring instead of condition (3) that the product
of any set of linearly free vectors in V should not belong to K. However,
even this would leave some 'ambiguity' in the definition by generators and
relations. The above definition results in a unique algebra only 'up to isomorphism'. Here are two more examples to clarify the meaning of this statement:

234

P. LOUNESTO

3. The multiplication table of the exterior algebra

basis 1, el, e2, el /\ e2 is
/\

o
o
o

Introduce a second product on

A
el
e2
el /\ e2

Arn. 2 with respect to the

Arn. 2 with multiplication table

el

e2
el /\ e2 + b

-el /\ e2 - b
-bel

-be2

el /\ e2
-bel
-be2
_b 2 - 2bel /\ e2

where b > o. Denote the second product by u Av. Rearrange the multiplication table of the second product into the form

A
el /\ e2 + b

o
o

0
0

which shows that we have generated a new exterior algebra Arn. 2 on rn. 2 ,
different from Arn. 2 but isomorphic with Arn. 2 In other words, we have
introduced a linear mapping a: Arn. 2 -+ Arn. 2 for which a(ed = ei, i =
1,2 and a( el /\ e2) = el Ae2 = el /\ e2 + b so that it is the identity on rn. 2
and gives an isomorphism between the two products a( u /\ v) = a( u) Aa( v).
4. An orthonormal basis el, e2 for rn. 2 satisfying eiej+ejei = 2liij generates
the Clifford algebra Cl 2 = Cl 2 ,o with basis 1, el, e2, e12 = el e2 (= el/\ e2).
We have the following multiplication table for Cl 2
el
e2
e12

el e2 e12
1 e12 e2
-e12 1 -el
-e2 el -1

Introduce a second product on Cl 2 with multiplication table

el
e2
e12

el
e2
el2
1
e2 - bel
e12 + b
-e12 - b
1
-el - be2
-e2 - bel el - be2 -1 - b2 - 2be12

The anticommutation relations eiej + ejei = 2liij are satisfied also by

the new product, and one may directly verify the associativity. As the real
number b varies we have a family of different but isomorphic Clifford algebras
on rn. 2

235

RIESZ'S CLIFFORD ALGEBRA

1.5.

UNIVERSAL OBJECT OF ALGEBRAS OF QUADRATIC FORMS

x 2 = Q(X)

The Clifford algebra Cl is the universal associative algebra over K generated

by V with the relations x 2 = Q(x), x E V.
Let Q be the quadratic form on a vector space V over a field K, and
let A be an associative algebra over K with unit lA. A linear mapping
V -+ A, x -+ If'x such that
(If'X)2

= Q(x) IA

for all x E V

is called a Clifford map. The sub algebra of A generated by K = K . lA

and V (or more precisely by the images of K and V in A) is called an
algebra of quadratic form Q. The Clifford algebra Cl is an algebra of
the quadratic form Q with a Clifford map V -+ Cl, x -+ IX such that for
any Clifford map cp: V -+ A there exists a unique algebra homomorphism
1/1 : Cl -+ A making the following diagram commutative

VLCl
cp'\.

11/1
A

This definition says that all Clifford maps may be obtained from I : V
which is thereby universal.

-+

Cl

The definition by the univeral property is meaningful for an algebraist who

knows categories and morphisms up to the theory of universal objects. A
category contains objects and morphisms between the objects. Invertible
morphisms are called isomorphisms. In a category there is an initial (resp.
final) universal object U, iffor any object A, there is a unique morphism a:
U -+ A (resp. A -+ U). The universal objects are unique up to isomorphism.
In many categories there exists trivially the final universal object, which
often reduces to o. Clifford algebra is the initial universal object in the
category of algebras of quadratic forms.

Example. Consider the category of algebras of the quadratic form on lRp,q

In this category the initial universal object is the Clifford algebra Clp,q of
dimension 2n and the final universal object is o. Between these two objects
there are no other objects, when p - q :f. 1 mod 4. However, there are
4 objects in this category, when p - q
1 mod 4; between Clp,q and 0
there are two algebras of both dimension 2n - 1 ; in one we have the relation
el e2 ... en = 1 and in the ohter el e2 ... en = -1; these two algebras are
not isomorphic in the category of algebras of quadratic forms (the identity
mapping on lRp,q does not extend to an isomorphism from one algebra to the
other); however, they are isomorphic as associative algebras (in the category

236

P. LOUNESTO

of all real algebras).

The above definition of Clifford algebras is most suitable for an algebraist

who wants to study Clifford algebras over commutative rings (and who does
not insist on injectivity of mappings K
A and V
A). However, this
approach does not guarantee the existence, which is given by constructing
the Clifford algebra as the quotient algebra of the tensor algebra (which in
turn is regarded by algebraists as the mother of all algebras).
1.6.

CLIFFORD ALGEBRA AS A QUOTIENT OF THE TENSOR ALGEBRA

V/I(Q)
Chevalley 1954 p. 37 constructs the Clifford algebra Ci(Q) as the quotient
algebra V/I(Q) of the tensor algebra V with respect to the two-sided
ideal I( Q) generated by the elements x x - Q(x) where x E V. See also
N. Bourbaki 1959 p.139 and T. Y. Lam 1973 p.l03. The tensor algebra approach gives a proof of existence by construction - suitable for an algebraist
who is interested in rapid access to the main properties of Clifford algebras.

* * * *
In characteristic zero we may avoid quotient structures by making the exterior algebra AV concrete as the subspace of antisymmetric tensors in V.
For example, if x,y E V, then x A y =
y - y x) E A2 V. More
generally, a simple k-vector Xl A X2 A ... A Xk is identified with
Alt(Xl X2 ... Xk)

;! E

sign(7r) x,..(l) X 1f(2) ... x,..(k),

1f

where the linear operator Alt : V

AV, called alternation, is a projection operator Alt(V) = AV satisfying u A v = Alt(u v). [Of course,
the subspace Alt( V) is not closed under the tensor product, though for
- B A) E A2 V.l
instance for A, BEN V we have A B - B A =
Similarly, we may obtain a linear isomorphism AV
Ci(Q) by identifying simple k-vectors with antisymmetric Clifford products

thus splitting the Clifford algebra Cl(Q) into fixed subspaces of k-vectors
Any orthogonal basis ell e2, ... , en of V gives a correspondence

Ak V c Cl( Q).

of bases for AV and Ci(Q). Some authors use unnecessarily complicated

tools in this context, such as a filtration of Cl( Q), needed only in characteristic 2, even though the above method does not allow division by 2.

RIESZ'S CLIFFORD ALGEBRA

237

2. Chapter 2: 'Rotations and Reflections'

3. Chapter 3: 'Canonical Representation of Isometries'
Chapter 2 on rotations and reflections and the Section 4.18 in Chapter 4 on
symmetric, absorbing, expelling and idempotent sets and norms have been
edited by John Horvath. Chapter 3 does not deal with Clifford algebras.
4. Chapter 4: 'Representation ofIsometries by Infinitesimal Transformations and Clifford Bivectors'
Among the four chapters of Riesz's monograph this last chapter has had the
greatest influence on physical applications of Clifford algebras. Chapter 4 is
a thorough investigation on squaring, exponentiation, decomposition, contraction and differentiation of bivectors in Clifford algebras. In particular,
Riesz condensed Maxwell's equations into a single equation with Clifford
bivectors (following here Juvet&Schidlof 1932 and A. Mercier 1935). However, to prevent this review from becoming too long I will stop here and
leave the reader the excitement of exploring the texts of Riesz.
Final Remark
Marcel Riesz's monograph is an endeavour to make Clifford algebras elementary and easily accessible to a wide public. This was also observed by
J. L. Tits in Mathematical Reviews 31 (1966) #6177. Tits focused on rigour,
though he observed that pedagogical aspects are the main concern of Riesz.
This renders Tits' review useless and unfair.
Acknowledgements
I would like to thank Ronald Shaw (Hull) for correcting my misconceptions
about quadratic forms in charactetistic 2 and Jacques Helmstetter (Grenoble) for teaching me the theory of universal objects and for emphasizing to
me the significance of the contraction.
Bibliography

In the text the years 1986 and 1992 refer to the Proceedings of the Canterbury (1985) and Montpellier (1989) workshops on Clifford algebras. For
historical references see the excellent surveys of Budinich&Trautman 1988,
Bolinder 1986, 1987 and van der Waerden 1985.
R. Ablamowicz, P. Lounesto, J. Maks: Conference Report, Second Workshop
on 'Clifford Algebras and Their Applications in Mathematical Physics,'

238

P. LOUNESTO

Universite des Sciences et Techniques du Languedoc, Montpellier, France,

1989. Found. Phys. 21 (1991),735-748.
L. Ahlfors, P. Lounesto: Some remarks on Clifford algebras. Complex Variables, Theory and Application 12 (1989), 201-209.
S. L. Altmann: Rotations, Quaternions and Double Groups. Clarendon Press,
Oxford, 1986.
E. Artin: Geometric Algebra. Interscience, New York, 1957, 1988.
M. F. Atiyah, R. Bott, A. Shapiro: Clifford modules. Topology 3, suppl. 1
(1964), 3-38. Reprinted in R. Bott: Lectures on K(X). Benjamin, New
York, 1969, pp. 143-178. Reprinted in Michael Atiyah: Collected Works,
Vol. 2. Clarendon Press, Oxford, 1988, pp. 301-336.
I. M. Benn, R. W. Tucker: An Introduction to Spinors and Geometry with
Applications in Physics. Adam Hilger, Bristol, 1987.
E. F. Bolinder: Unified microwave network theory based on Clifford algebra
in Lorentz space. Conference Proceedings, 12th European Microwave Conference (Helsinki 1982), Microwave Exhibitions and Publishers, Turnbridge
Wells, 1982, pp. 25-35.
E.F. Bolinder: Clifford algebra: What is it? IEEE Antennas and Propagation Society Newsletter 29 (1987), 18-23.
N. Bourbaki: Algebre, Chapitre 9, Formes sesquilineaires et formes quadratiques. Hermann, Paris, 1959.
F. Brackx, R. Delanghe, F. Sommen: Clifford Analysis. Research Notes in
Mathematics 16, Pitman Books, London, 1982.
R. Brauer, H. Weyl: Spinors in n dimensions. Amer. J. Math. 51 (1935),
425-449. Reprinted in Selecta Hermann Weyl, Birkhiiuser, Basel, 1956, pp.
431-454.
P. Budinich, A. Trautman: The Spinorial Chessboard. Springer, Berlin, 1988.
R. Carmichael: Book Review (F. Brackx, R. Delanghe, F. Sommen: Clifford
Analysis). Bull. Amer. Math. Soc. 11 (1984),227-240.
E. Cartan (expose d'apres l'article allemand de E. Study): Nombres complexes. In J. Molk (red.): Encyclopedie des sciences mathematiques, Tome
I, vol. 1, Fasc. 4, art. IS (1908), pp. 329-468. Reprinted in E. Cartan:
CEuvres Completes, Partie II. Cauthier-Villars, Paris, 1953, pp. 107-246.
C. Chevalley: Theory of Lie Groups. Princeton University Press, Princeton,
1946.
C. Chevalley: The Algebraic Theory of Spinors. Columbia University Press,
New York, 1954.
C. Chevalley: The Construction and Study of Certain Important Algebras.
Mathematical Society of Japan, Tokyo, 1955.
J. S.R. Chisholm, A.K. Common (eds.): Proceedings of the NATO and SERC
Workshop on 'Clifford Algebras and Their Applications in Mathematical
Physics,' Canterbury, England, U.K., 1985. Reidel, Dordrecht, 1986.
W.K. Clifford: Applications of Grassmann's extensive algebra. Amer. J.

RIESZ'S CLIFFORD ALGEBRA

239

Math. 1 (1878), 350-358.

W. K. Clifford: On the classification of geometric algebras. In R. Tucker (ed.):
Mathematical Papers by William Kingdon Clifford, Macmillan, London,
1882, pp. 397-401. (Reprinted by Chelsea, New York, 1968.) Title of talk
announced already in Proc. London Math. Soc. 7 (1876), p. 135.
A. Crumeyrolle: Orthogonal and Symplectic Clifford Algebras, Spinor Structures. Kluwer, Dordrecht, 1990.
R. Deheuvels: Formes quadratiques et groupes classiques. Presses Universitaires de France, Paris, 1981.
R. Delanghe, F. Sommen, V. Soucek: Clifford Algebra and Spinor Valued
Functions: A Function Theory for the Dirac Operator. Kluwer, Dordrecht,
1992.
P. A. M. Dirac: The quantum theory of the electron. Proc. Roy. Soc. A117
(1928), 610-624.
L. Gfu-ding: Marcel Riesz in memoriam. Acta Math. 124 (1970), pp. I-XI.
Reprinted in M. Riesz, Collected Papers, pp. 1-9.
L. Gci.rding, L. Hormander (eds.): Marcel Riesz, Collected Papers. Springer,
Berlin, 1988.
J. Gilbert, M. Murray: Clifford Algebras and Dirac Operators in Harmonic
Analysis. Cambridge Studies in Advanced Mathematics, Cambridge University Press, Cambridge, 1991.
W. Greub: Multilinear Algebra, 2nd edn. Springer, Berlin, 1978.
F. R. Harvey: Spinors and Calibrations. Academic Press, San Diego, 1990.
J. Helmstetter: Algebres de Clifford et algebres de Weyl. Cahiers Math. 25,
Montpellier, 1982.
D. Hestenes: Space-Time Algebra. Gordon and Breach, New York, 1966,
1987, 1992.
D. Hestenes, G. Sobczyk: Clifford Algebra to Geometric Calculus. Reidel,
Dordrecht, 1984, 1987.
B. Jancewicz: Multivectors and Clifford Algebra in Electrodynamics. World
Scientific Publ., Singapore, 1988.
G. Juvet: Operateurs de Dirac et equations de Maxwell. Comment. Math.
Helv. 2 (1930),225-235.
E. Kahler: Der innere Differentialkalkiil. Rendiconti di Matematica e delle
sue Applicazioni (Roma) 21 (1962), 425-523.
Y. Kawada, N. Iwahori: On the structure and representations of Clifford
algebras. J. Math. Soc. Japan 2 (1950),34-43.
M.-A. Knus: Quadratic Forms, Clifford Algebras and Spinors. Univ. Estadual de Campinas, SP, 1988.
T. Y. Lam: The Algebraic Theory of Quadratic Forms. Benjamin, Reading,
MA, 1973, 1980.
O. Laporte, G. E. Uhlenbeck: Application of spinor analysis to the Maxwell
and Dirac equations. Phys. Rev. 37 (1931), 1380-1397.

240

P. LOUNESTO

H. B. Lawson, M.-L. Michelsohn: Spin Geometry. Princeton Univerity Press,

Princeton, NJ, 1989.
H. C. Lee: On Clifford algebras and their representations. Ann. of Math. 49
(1948),760-773.
R. Lipschitz: Principes d'un calcul algebrique qui contient comme especes
particulieres Ie calcul des quantites imaginaires et des quaternions. C. R.
Acad. Sci. Paris 91 (1880),619-621,660-664. Reprinted in Bull. Soc. Math.
(2) 11 (1887), 115-120.
R. Lipschitz: Untersuchungen iiber die Summen von Quadraten. Max Cohen
und Sohn, Bonn, 1886, pp. 1-147. (The first chapter ofpp. 5-57 translated
into French by J. Molk: Recherches sur la transformation, par des substitutions reelles, d 'une somme de deuz ou troiz carres en elle-meme. J. Math.
Pures Appl. (4) 2 (1886),373-439. French resume of all three chapters in
Bull. Sci. Math. (2) 10 (1886), 163-183.
R. Lipschitz (signed): Correspondence. Ann. of Math. 69 (1959),247-251.
P. Lounesto: Report of Conference, NATO and SERC Workshop on 'Clifford
Algebras and Their Applications in Mathematical Physics,' University of
Kent, Canterbury, England, 1985. Found. Phys. 16 (1986),967-971.
A. Micali, Ph. Revoy: Modules quadratiques. Cahiers Math. 10, Montpellier,
1977. Bull. Soc. Math. France 63, suppl. (1979),5-144.
A. Micali, R. Boudet, J. Helmstetter (eds.): Proceedings of the 2nd Workshop on 'Clifford Algebras and their Applications in Mathematical Physics,'
Montpellier, France, 1989. Kluwer, Dordrecht, 1992.
W. Pauli: Zur Quantenmechanik des magnetischen Elektrons. Z. Phys. 42
(1927), 601-623.
I. R. Porteous: Topological Geometry. Van Nostrand Reinhold, London, 1969.
Cambridge University Press, Cambridge, 1981.
M. Riesz: En ciskQ.dlig bild av den icke-euklidiska geometrien, geometriska
strovtcig inom relativitetsteorien. Lunds Universitets Arsskrift, Lund, 1943,
pp.3-76.
M. Riesz: L'integrale de Riemann-Liouville et Ie probleme de Cauchy. Acta
Math. 81 (1949),1-123; Collected Papers, pp. 571-793.
M. Riesz: Sur certain notions fondamentales en theorie quantique relativiste.
C. R. 10e Congres Math. Scandinaves, Copenhagen, 1946. Jul. Gjellerups
Forlag, Copenhagen, 1947, pp. 123-148; Collected Papers, pp. 545-570.
M. Riesz: L 'equation de Dirac en relativite generale. Tolfte( =12.) Skandinaviska Matematikerkongressen i Lund, 1953. Hakan Ohlssons boktryckeri,
Lund, 1954, pp. 241-259; Collected Papers, pp. 814-832.
M. Riesz: Clifford Numbers and Spinors. The Institute for Fluid Dynamics
and Applied Mathematics, Lecture Series No. 38, University of Maryland,
1958.
S. Salamon: Riemannian Geometry and Holonomy Groups. Longman Scientific, Essex, 1989.

RIESZ'S CLIFFORD ALGEBRA

241

F. Sauter: Losung der Diracschen Gleichungen ohne Spezialisierung der

Diracschen Operatoren. Z. Phys. 63 (1930),803-814.
L. Silberstein: The Theory of Relativity. Macmillan, London, 1914.
K. Th. Vahlen: tiber hohere komplexe Zahlen. Sehriften der phys.-okon.
Gesellschaft zu Konigsberg 38 (1897),72-78.
K. Th. Vahlen: tiber Bewegungen und complexe Zahlen. Math. Ann. 55
(1902), 585-593.
B. L. van der Waerden: On Clifford algebras. Nederl. Akad. Wetensch. Proe.
Ser. A69 (1966), 78-83.
B. L. van der Waerden: A History of Algebra. Springer, Berlin, 1985.
E. Witt: Theone der quadratischen Formen in beliebigen Korpern. J. Reine
Angew. Math. 176 (1937),31-44.