Simplied production of

DIRAC DELTA FUNCTION IDENTITIES

Nicholas Wheeler, Reed College Physics Department
November 1997
Introduction. To describe the smooth distribution of (say) a unit mass on the
r-axis, we introduce distribution function j(r) with the understanding that
j(r) dr mass element d: in the neighborhood dr of the point r
_
j(r) dr = 1
To describe a mass distribution localized to the vicinity of r = o we might, for
example, write
j(r o; c) =
_

_
1
2
if o c < r < o + c, and 0 otherwise; else
1

2
exp
_

1
2
(r o)
2
_
; else
1
x
sin(r,c); else . . .
In each of those cases we have
_
j(r o; c) dr = 1 for all c 0, and in each
case it makes formal sense to suppose that
lim
0
j(r o; c) describes a unit point mass situated at r = o
Dirac clearly had precisely such ideas in mind when, in 15 of his Quantum
Mechanics,
1
he introduced the point-distribution (r o). He was well aware
1
I work from his Revised 4
th
Edition (16), but the text is unchanged from
the 3
rd
Edition (1). Diracs rst use of the -function occurred in a paper
published in 1:6, where (rj) was intended to serve as a continuous analog
of the Kronecker delta
mn
, and thus to permit unied discussion of discrete
and continuous spectra.
2 Simplied Dirac identities
that the delta functionwhich he presumes to satisfy the conditions
_
+

(r o) dr = 0
(r o) = 0 for r ,= o
is not a function. . . according to the usual mathematical denition; it is,
in his terminology, an improper function, a notational device intended to
by-pass distracting circumlocutions, the use of which must be conned to
certain simple types of expression for which it is obvious that no inconsistency
can arise.
Diracs cautionary remarks (and the ecient simplicity of his idea)
notwithstanding, some mathematically well-bred people did from the outset
take strong exception to the -function. In the vanguard of this group was
John von Neumann, who dismissed the -function as a ction, and wrote
his monumental Mathematische Grundlagen der Quantenmechanik
2
largely to
demonstrate that quantum mechanics can (with sucient eort!) be formulated
in such a way as to make no reference to such a ction.
The situation changed, however, in 1o, when Laurent Schwartz published
the rst volume of his demanding multi-volume Theorie des distributions.
Schwartz accomplishment was to show that -functions are (not functions,
either proper or improper, but) mathematical objects of a fundamentally new
typedistributions, that live always in the shade of an implied integral sign.
This was comforting news for the physicists who had by then been contentedly
using -functions for thirty years. But it was news without major consequence,
for Schwartz work remained inaccessible to all but the most determined of
mathematical physicists.
Thus there came into being a tradition of simplication and popularization.
In 1 Schwartz gave a series of lectures at the Seminar of the Canadian
Mathematical Congress (held in Vancouver, B.C.), which gave rise in 1: to
a pamphlet
3
that circulated widely, and brought at least the essential elements
of the theory of distributions into such clear focus as to serve the simple needs
of non-specialists. In 1 the British applied mathematician G. Temple
building upon remarks published a few years earlier by Mikusnski
4
published
what he called a less cumbersome vulgarization of Schwartz theory, which
he hoped might better serve the practical needs of engineers and physicists.
Temples lucid paper inspired M. J. Lighthill to write the monograph from
which many of the more recent introductions to the theory of distributions
2
The German edition appeared in 1:. I work from the English translation
of 1. Remarks concerning the -function can be found in 3 of Chapter I.
3
I. Halperin, Introduction to the Theory of Distributions.
4
J. G. Mikusnski, Sur la methode de generalization de Laurent Schwartz
et sur la convergence faible, Fundamenta Mathematicae 35, 235 (1948).
Introduction 3
descend. Lighthills slender volume
5
by intention a text for undergraduates
bears this dedication
to
paul dirac
who saw that it must be true
laurent schwartz
who proved it,
and
george temple
who showed how simple it could be made
and has about itas its title promisesa distinctly Fourier analytic avor.
Nor is this fact particularly surprising; the Fourier integral theorem
)(r) =
1
2
_
+

c
ikx
__
+

c
+iky
)(j) dj
_
d/ for nice functions )()
can by reorganization be read as an assertion that
1
2
_
+

c
ik(yx)
d/ = (j r)
The history of the -function can in this sense be traced back to the early
18:os. Fourier, of course, was concerned with the theory of heat conduction,
but by 18:8 the -function had intruded for a second time into a physical theory;
George Green noticed that the solution of the Poisson equation
2
(r) = (r),
considered to describe the electrostatic potential generated by a given charge
distribution (r), can be obtained by superposition of the potentials generated
by a population of point charges; i.e., that the general problem can be reduced
to the special problem

2
(r; j) = (r j)
where now the -function is being used to describe a unit point charge
positioned at the point j. Thus came into being the theory of Greens
functions, whichwith important input by Kirchho (physical optics, in the
188os) and Heaviside (transmission lines, in the 18os)became, as it remains,
one of the principal consumers of applied distribution theory.
I have sketched this history
6
in order to make clear that what I propose
to do in these pages stands quite apart, both in spirit and by intent, from the
trend of recent developments, and is fashioned from much ruder fabric. My
objective is to promote a point of viewa computational techniquethat came
5
Introduction to Fourier Analysis & Generalized Functions (18).
6
Of which Jesper L utzen, in his absorbing The Prehistory of the Theory of
Distributions (18:), provides a wonderfully detailed account. In his Concluding
Remarks L utzen provides a nicely balanced account of the relative contributions
of Schwartz and of S. L. Sobelev (in the early 1os).
4 Simplied Dirac identities
accidentally to my attention in the course of work having to do with the one-
dimensional theory of waves.
7
I proceed very informally, and will be concerned
not at all with precise characterization of the conditions under which the things
I have to say may be true; this fact in itself serves to separate me from recent
tradition in the eld.
Regarding my specic objectives. . . Dirac remarks that There are a
number of elementary equations which one can write down about functions.
These equations are essentially rules of manipulation for algebraic work
involving functions. The meaning of any of these equations is that its two
sides give equivalent results [when used] as factors in an integrand. Examples
of such equations are
(r) = (r)
r(r) = 0
(or) = o
1
(r) : o 0 (1.1)
(r
2
o
2
) =
1
2
o
1
_
(r o) + (r + o)
_
: o 0 (1.2)
_
(o r) dr(r /) = (o /)
)(r)(r o) = )(o)(r o)
On the evidence of this list (which attains the length quoted only in the
3
rd
edition) L utzen concludes that Dirac was a skillful manipulator of the
-function, and goes on to observe that some of the above theorems, especially
(1.2), are not even obvious in distribution theory, since the changes of variables
are hard to perform. . .
8
The formal identities in Diracs list are of several
distinct types; he supplies an outline of the supporting argument in all cases
but one: concerning (1) he remarks only that they may be veried by similar
elementary arguments. But the elementary argument that makes such easy
work of (1.1)
9
acquires a fussy aspect when applied to expressions of the form

_
p(r)
_
typied by the left side of (1.2). My initial objective will be to
demonstrate that certain kinds of -identities (including particularly those of
type (1)) become trivialities when thought of as corollaries of their -analogs. By
extension of the method, I will then derive relationships among the derivative
properties of () which are important to the theory of Greens functions.
7
See R. Platais, An investigation of the acoustics of the ute (Reed College
physics thesis, 1).
8
See Chapter 4, 29 in the monograph previously cited.
9
By change of variables we have
_
)(r)(or) dr =
_
)(j,o)(j)
1
|a|
dj =
1
|a|
)(0)
=
_
)(j)
1
|a|
(j) dj
which assumes only that the Jacobian [o[ , = 0.
Heaviside step function 5
1. Properties and applications of the Heaviside step function.The step function
()introduced by Heaviside to model the action of a simple switchcan be
dened
(r) =
_
_
_
0 for r < 0
1
2
at r = 0
1 for r 0
(2)
where the central
1
2
is a (usually inconsequential) formal detail, equivalent to
the stipulation that
(r) 2(r) 1 =
_
_
_
1 for r < 0
0 at r = 0
+1 for r 0
(3)
be odd (i.e., that () vanish at the origin). As Dirac himself (and before him
Heaviside) have remarked, the step function (with which Dirac surely became
acquainted as a student of electrical engineering) and the -function stand in a
close relationship supplied by the calculus:
(r o) =
_
x

(j o) dj (4.1)

d
dx
(r o) = (r o) (4.2)
The central
1
2
is, in this light, equivalent to the stipulation that (r) be
(formally) an even function of r. For the same reason that (r) becomes
meaningful only in the shade of an integral sign, so also does (r), at least
as it is used in intended applications; the construction
(r) lim
0
(r; c) entails (r) = lim
0
_
(r; c)
_
x

(j; c) dj
_
and causes (r; c) to become literally dierentiable at the origin, except in the
limit.
That (r) and (r) are complementary constructs can be seen in yet
another way. The identity
)(r) =
_
+

(r j))(j) dj (5)
provides what might be called the picket fence representation of )(r). But
=
_
+

d
dy
(r j)
_
)(j) dj
=
_
+

(r j))

(j) dj + boundary term (6)

which (under conditions that cause the boundary term to vanish) provides the
less frequently encountered stacked slab representation of )(r). In the former
it is )() itself that serves to regulate the heights of successive pickets, while
in the latter it is not ) but its derivative )

() that regulates the thicknesses

of successive slabs. For graphical representations of (5) and (6) see Figure 1.
6 Simplied Dirac identities
Figure 1: The picket fence representation (5) of )(r), compared
with the stacked slab representation (6).
Partial integration (subject always to the presumption that boundary terms
vanish), which we used to obtain (6), is standardly used also to assign meaning
to the successive derivatives of the -function; one writes
_
)(j)

(j r) dj =
_
)

(j)(j r) dj = )

(r)
_
)(j)

(j r) dj =
_
)

(j)

(j r) dj
= ()
2
_
)

(j)(j r) dj = ()
2
)

(r)
.
.
.
_
)(j)
(n)
(j r) dj = ()
n
)
(n)
(r)
Formally, one has (see Figure 2)

(j r) = lim
0

_
j (r c)
_

_
j (r + c)
_
2c

(j r) = lim
0

_
j (r 2c)
_
2
_
j r) +
_
j (r + 2c)
__
(2c)
2
.
.
.
_

_
(7)
which when introduced as factors into the integrand of
_
)(j) dj serve simply
to reconstruct the denitions of )

(r), )

(r), etc.
In anticipation of future needs, I note that the preceding equations can
be formulated as relations among -functions. Taking (4.2) as our point of
departure, we have
(j r) = lim
0

_
j (r c)
_

_
j (r + c)
_
2c
= limit of a sequence of ever-narrower/taller box functions
Simplied derivation of delta function identities 7
x
y
x
Figure 2: The gures on the left derive from (7), and show
representations of ascending derivatives of (j r). The gures
on the right derive from (8), and provide representations of the
same material.
giving (see again the preceding gure)

(j r) = lim
0
(y[x2])2(yx)+(y[x+2])
(2)
2

(j r) = lim
0
(y[x3])3(y[x])+3(y[x+])(y[x+3])
(2)
3
.
.
.
_

_
(8)
2. Simplied derivation of delta function identities. Let (r; c) refer to some
(any nice) parameterized sequence of functions convergent to (r), and let o
be a positive constant. While (or; c) and (r; c) are distinct functions of
r, they clearly become identical in the limit c 0, and so also therefore do
their derivatives (of all orders). So we have o

(or) =

(r), which by (4.2)

reproduces (1.1):
(or) = o
1
(r) : o 0 (9.1)
8 Simplied Dirac identities
If, on the other hand, o < 0 then (or) = 1 (r) gives o

(or) =

(r)
whence
(or) = o
1
(r) : o < 0 (9.2)
The equation
(or) =
1
|a|
(r) : o ,= 0 (10.1)
provides a unied formulation of (9.1) and (9.2). A second dierentiation gives
o
2

(or) =

(r) according as o 0

(or) =
1
a
2

(r) (10.2)
Preceding remarks illustrate the sense in which -identities become trivialities
when thought of as corollaries of their -analogs. By way of more interesting
illustration of the same point. . .
The function p(r) r
2
o
2
describes an up-turned parabola that crosses
the r-axis at r = o. Evidently
(r
2
o
2
) =
_
_
_
1 for r < o
0 for o < r < +o
1 for r +o
= 1
_
(r + o) (r o)
_
by dierentiation entails
2r(r
2
o
2
) = (r + o) + (r o)

(r
2
o
2
) = (2o)
1
(r + o) + (+2o)
1
(r o)
=
1
2|a|
_
(r o) + (r + o)
_
(11)
which provides a generalized formulation if (1.2). By this mode of argument it
becomes transparently clear how the o that enters into the prefactor comes to
acquire its (otherwise perplexing) absolute value braces.
Suppose, more generally, that
p(r) = p
0
(r r
1
)(r r
2
) (r r
n
)
with r
1
< r
2
< < r
n
. We then (see Figure 3) have
(p(r)) =
_
_
_
(r r
1
) (r r
2
) + ()
n
(r r
n
)
1
_
(r r
1
) (r r
2
) + ()
n
(r r
n
)
_
(12)
Elementary geometrical theory of Greens functions 9
Figure 3: Representation of (p(r)) in a typical case.
according as p

(r
1
) 0. By dierentiation
p

(r) (p(r)) =
_
(r r
1
) (r r
2
) + ()
n
(r r
n
)
_

(p(r)) =
n

k=1
1
|g

(x
k
)|
(r r
k
) in all cases (13)
The identities (10.1) and (11) are seen to comprise special instances of this
famous result. The more standard derivation involves massaging the integral
_
)(r)(p(r)) dr, and is a good deal less immediate. The approach here
advocated owes its simplicity very largely to the simplicity of (12).
3. Elementary geometrical approach to properties of some Greens functions. It
was known already to dAlembert (1) that if (t, r) is a solution of the wave
equation
10
= 0 with
2
t

2
x
and if initial data (0, r) and
t
(0, r) are prescribed, then
(t, r) =
1
2
_
(0, r t) + (0, r + t)
_
+
1
2
_
x+t
xt

t
(0, j) dj (14)
= function of (r t) + function of (r + t)
= right-running wave + left-running wave
10
For the purposes of this discussion I nd it convenient (though alien to my
practice) to set c = 1.
10 Simplied Dirac identities
But two centuries were to elapse before it came to be widely appreciated that
(14) can usefully be notated
(t, r) =
_
+

0
(r j, t)(0, j) +
0
(r j, t)
t
(0, j)
_
dj (15)
with

0
(r j, t)
1
2
_

_
j (r t)
_

_
j (r + t)
_
_
(16.1)

0
(r j, t)
1
2
_

_
j (r t)
_
+
_
j (r + t)
_
_
(16.2)
=

t

0
(r j, t)
It is important to notice that, while we had prediction (i.e., evolution from
prescribed initial data) in mind when we devised (15), the equation also works
retrodictively; (15) is invariant under t t. Under time-reversal all
t
s
change sign, but so also (manifestly) does
0
(rj, t), and when we return with
this information to (15) we see that the two eects precisely cancel. Evidently

0
(rj, t n)thought of as a function of n, j that depends parametrically
on t, rhas the properties

0
(r j, t n) =
_

_
+
1
2
inside the lightcone extending backward from t, r

1
2
inside the lightcone extending forward from t, r
0 outside the lightcone
=
1
2
_

_
(j r) + (t n)
_

_
(j r) (t n)
_
_
=
1
2
(t n)
_
(t n)
2
(r j)
2
_
illustrated in Figure 4. In order for the right side of (15) to do its assigned work
it must be the case that
_

2
t

2
x
_

0
(r j, t n) = 0
and

0
(r j, 0) = 0

0
t
(r j, 0) = (j r)

0
tt
(r j, 0) = 0
all of which are, in fact, quick consequences of (16).
The forced wave equation reads
= 1 with 1(t, r) prescribed
The general solution can be developed
=
0
+
F
where
0
evolves according to
0
= 0 from prescribed initial data, while
F
evolves according to
F
= 1 but is assumed to vanish on the initial timeslice.
Elementary geometrical theory of Greens functions 11
t
x
u
y
+
_
Figure 4: Representation of the Greens function
0
(rj, tn) of
the homogeneous wave equation = 0. For n < t the function has
the form of a triangular plateau (backward lightcone) with a at top
at elevation
1
2
, while for n t (forward lightcone) it is a triangular
excavation of similar design.
Greens method leads one to write

F
(t, r) =
_ _
(r j, t n)1(n, j) djdn
and to require of (r j, t n) that
_

2
t

2
x
_
(r j, t n) = (j r)(n t) (17)
and (j r, 0) = 0. Such properties are possessed by (in particular) this close
relative of
0
(r j, t n) = (t n)
0
(r j, t n) (18)
=
_
1
2
_

_
(j r) + (t n)
_

_
(j r) (t n)
_
: t n 0
0 : t n < 0
=
_
_
_
+
1
2
inside the lightcone extending backward from t, r
0 inside the lightcone extending forward from t, r
0 outside the lightcone
12 Simplied Dirac identities
which is known as the retarded propagator. The properties in question
are possessed also by the similarly-dened advanced propagator, and are
therefore shared also by all functions of the form
11
(, ) = /
ret
(, ) + (1 /)
adv
(, )
We encounter at this point the need for a principle of choice, but it would
distract me from my primary objective to pursue the matter.
Concerning that primary objective: I entered into the preceding review
of what is after all standard material partly to demonstrate that the Fourier
analytic methods standardly encountered
12
are, in fact, inessential (the subject
is susceptible to development by elementary means). . . but mainly to place
Figure 5: Representation of the truncation process (18), by
means of which
0
gives rise to .
myself in position to pose this question: What is the detailed mechanism by
means of which truncation causes a solution
0
of the homogeneous wave
equation to become a solution of the inhomogeneous equation:

0
= 0
truncation of
0
=
Elementary insight into the geometrical essence of the answer is provided by
the following sequence of gures:
11
See F. Rohrlich, Classical Charged Particles (16), p. 79.
12
See, for example, relativistic classical fields (1), pp. 163190 and
analytical methods of physics (181), pp. 291297.
Elementary geometrical theory of Greens functions 13
-1
0
+1
-1
t
+1
x
Figure 6: The numbers, when multiplied by
1
2
(2c)
2
, referin the
sense familiar from (7) and (8); see also the lower right detail in
Figure 2to the discrete approximation
+
2
t
(r j, t n)
+
(xy,[t+2]u)2(xy,[t]u)+(xy,[t2]u)
(2)
2
In this and subsequent gures (r j, t n) is considered to be a
function of n, j, into which t, r enter as parameters.
14 Simplied Dirac identities
+1
-1
0
+1
-1
t
x
+2
Figure 7: The numbers, when multiplied by
1
2
(2c)
2
, refer to the
discrete approximation

2
x
(r j, t n)

([x+2]y,tu)2([x]y,tu)+([x2]y,tu)
(2)
2
Elementary geometrical theory of Greens functions 15
0
+1
t
x
+1
0
0
0
0
Figure 8: Superposition of the preceding gures. The numbers,
when multiplied by
1
2
(2c)
2
, supply the discrete approximation
_

2
t

2
x
_
(r j, t n)
1
2
(2c)
2
Box(t, r; c)
where
Box(t, r; c)
_
1 interior to shaded box centered at (t, r)
0 elsewhere
The box has diagonal measure 4c, and its area is given therefore by
(4c,

2)
2
= 8c
2
. It is the top of a little prism of height
1
2
(2c)
2
, the
volume of which is given by
volume = 1 (all values of c)
evidently
lim
0
1
2
(2c)
2
Box(t, r; c) = (n t)(j r)
which yields the result we sought to establish.
16 Simplied Dirac identities
4. Analytical derivation of the same result. The discussion in the preceding
section culminated in a geometrical argument designed to illuminate how it
comes about that

0
= 0 but truncation engenders =
For comparative purposes I now outline the analytical demonstration of that
same fact; the details are not without interest, but the argument as a whole
seems to me to lack the ah-ha! quality of its geometrical counterpart.
By (18) we have
(r j, t n) =
1
2
(t n)(&c.)
&c. (t n)
2
(r j)
2
Therefore

t
=
1
2
(t n)(&c.) + (t n)(t n)(&c.)

t
=
1
2

(t n)(&c.) + 2(t n)(t n)(&c.)

+ (t n)(&c.) + (t n)2(t n)
2

(&c.)

x
= (t n)(r j)(&c.)

x
= (t n)(&c.) + (t n)2(r j)
2

(&c.)
giving
_

2
t

2
x
_
=
1
2

(t n)(&c.) + 2(t n)(t n)(&c.)

+ 2(t n)
_
(&c.) + (&c.)

(&c.)
_
. .
0
where the term that vanishes does so because r(r) = 0 = (r) +r

(r) = 0.
We note in passing that had we omitted the -factor from the denition of ;
i.e., if we were evaluating
0
instead of , we would at this point have
achieved
0
= 0.
Our assignment now is to establish that
1
2

(t n)(&c.) + 2(t n)(t n)(&c.) = (t n)(r j)

which, if we write t n and r j, can be notated
1
2

()(
2

2
) + 2()(
2

2
)
. .
= ()()
But
=
1
2

_
()(
2

2
)
_
. .
+()(
2

2
)
I will argue that
= 0 (19)
Dimensional generalization 17
Assuming, for the moment, the truth of that claim, we want to show that
() (
2

2
)
. .
= ()() ()
But this is in fact immediate; looking to (11) and taking advantage of
simplications made available by the presence of the ()-factor, we have
= ()
1
2
_
( ) + ( + )
_
= ()
1
2
_
() + ()
_
= () ()
Returning now to the demonstration of (19); if 1() is nice function, then
(integrating by parts) we have
_

_
()(
2

2
)
_
1() d =
_
()(
2

2
)1

() d
=
_
()
_
1
_
( + ) ( )
_
1

() d by (12)
=
_
()1

() d +
_
+

()1

() d
= 0 all 0, also in the limit 0 because 1() is nice
The argument just concluded is notable for its delicacy
13
and its overall
improvisatory, ad hoc quality. It is, in my view, not only less illuminating but
also less convincing than the geometrical argument summarized in Figures 68.
It points up the need for a systematic calculus of distributions, and lends
force to Diracs observation that we should (in the absence of such a calculus)
avail ourselves of -methods only when it is obvious that no inconsistency can
arise. This, however, is more easily said than done.
5. Relaxation of the assumption that spacetime is 2-dimensional. In 3 we
were able to obtain Greens functions of the homogeneous/inhomogeneous wave
equations
_

2
t

2
x
_

0
= 0
_

2
t

2
x
_
=
by putting dAlembert and Heaviside/Dirac in a pot and stirring gently, the
notable fact being that no Fourier was called for by the recipe. dAlembert made
critical use, however, of circumstances which are special to the 2-dimensional
13
We would, for example, have gone o-track if we had made too-casual
use of Diracs identity r(r) = 0; this makes good sense in the context Dirac
intended (
_
r(r) )(r) dr = 0), but can lead to error when (as at ()) r(r)
appears as a factor in a more complex expression.
18 Simplied Dirac identities
case; in higher-dimensional cases dAlemberts line of argument breaks down,
and new analytical methods are called for. Those methodsor at least the
methods most commonly brought to bear upon the analysis of
_

2
t

2
x
1
2
x
2
2
x
n
_

0
n
(r) = 0
_

2
t

2
x
1
2
x
2
2
x
n
_

n
(r) = (r)
make essential use of Fourier transform and contour integral techniques. The
small point to which I wish here to draw attention is that it is nevertheless
possible to omit Fourier from the recipe in the general case; the simple case
n = 1 provides the seed from which results appropriate to the general case can
be obtained by elementary operations borrowed from the so-called fractional
calculus. Of course, any person with practical need of (say)
0
n
(r) is a person
certain to be comfortable with Fourier analytic methods, grateful for the special
insights which those methods typically provide. . . and (in all likelihood) ignorant
of the concept of semi-dierentiation. Such a person may be inclined to
dismiss the following remarks as an amusing tour de force. To do so would,
however, be to incur the wrath of the ghosts of Jacques Hadamard and Marcel
Riesz, who originally undertook this work not to be cute but to clarify certain
deep problems.
One proceeds from the observation
14
that the variables t, r
1
, r
2
, . . . , r
n

enter into the structure of

0
n
(r) only in lumped combination
= t
2
:
2
with :
2
r
2
1
+ r
2
2
+ + r
2
n
and by fairly straightforward argument is led at length to the scheme

1

3

5

7
. . . etc.

2

4

6
where horizontal arrows refer to the action of
1

=
1
2r

r
and sloping
arrows refer to the action of the semi-dierentiation operator
_
1

_1
2
. The
gure that results when all arrows are reversed makes complementary good
sense, provided the reversed arrows are understood to refer to (semi)integration
processes. The scheme is computationally very ecient, and it exposes patterns
that tend to remain concealed when one proceeds by Fourier analytic means.
In particular, it makes evident (when one looks to the details) the profound
sense in which physics in odd-dimensional space is to be distinguished from
physics in even-dimensional space. And the sensevisible only when one is in
position to make comparative statementsin which the case n = 3 is special.
We have encountered a situation in which (arguably) simpler is also deeper.
And (which is more immediately germane to the business at hand) we have
encountered a context within which not only and but also

. . . enter
directly and essentially into the description of the physics.
14
See 7 of my Construction & physical application of the fractional
calculus (1) for mathematical details and physical commentary.