The First Systems of

Weighted
Differential and Integral

Calculus

Jane Grossman
Michael Grossman
Robert Katz
 THE FIRST SYSTEMS OF

WEI G HT ED

DIFFERENTIAL AND INTEGRAL

CALCULUS

Jane Grossman
University of Lowell
Michael Grossman
University of Lowell
Robert Katz
Arahimedes Foundation

Archimedes Foundation
Box 240, Rockport
Massachusetts 01966

Thi s One

3BJ3-F4Q-JQPO
"In general the position as regards all such
new calculi is this - That one cannot accom
plish by them anything that could not be
accomplished without them. However, the ad
vantage is, that, provided such a calculus
corresponds to the inmost nature of frequent
needs, anyone who masters it thoroughly is
able - without the unconscious inspiration
of genius which no one can command - to solve
the respective problems, yea, to solve them
mechanically in complicated cases in which,
without such aid, even genius becomes power
less. Such is the case with the invention of
general algebra, with the differential calcu
lus, and in a more limited, region with Lagran-
ge ' s calculus of variations, with my calculus
of congruences, and with Mobius's calculus.
Such conceptions unite, as it were, into an
organic whole countless problems which other
wise would remain isolated and require for
their separate solution more or less applica
tion of inventive genius."

Gauss

First Printing, July 1980

0-9771170-1-4
 PREFACE

Mathematicians have long known about weighted averages,

Stieltjes integrals, and derivatives of one function with res
pect to another function. But apparently no one has hitherto
noticed that by means of suitably modified definitions those
three concepts can be packaged into systems of calculi, com
plete with two fundamental theorems, mean value theorems, and
analogues of other important theorems of classical calculus.
We refer to the new systems as weighted calculi, because in
each such system a weight function plays a central role.

In Chapter 1 we present the classical calculus in a no

vel manner that leads naturally to the subsequent construction
of the weighted calculi.

Chapter 2 contains a development of the weighted clas

sical calculi, which involve weighted arithmetic averages,
weighted classical integrals (these are Stieltjes integrals) ,
and weighted classical derivatives (these are classical deri
vatives of one function with respect to another function) .
We were surprised to discover that many of the weighted clas
sical calculi are (unweighted) non-Newtonian calculi. (The
non-Newtonian calculi were discovered by Michael Grossman and
Robert Katz and discussed in their book Non-Newtonian Calculus ,
which was published by Lee Press in 1972. A familiarity with
that book is presupposed only for Section 2.ll and Chapter 3.)

Chapter 3 includes a development of the weighted

non-Newtonian calculi, which involve weighted non-Newtonian
averages, weighted non-Newtonian integrals, and weighted non-
Newtonian derivatives. The chapter concludes with a brief
presentation of weighted geometric calculus, an interesting
example of a weighted non-Newtonian calculus.

111
 Since this book is intended for a wide audience, includ
ing students, engineers, scientists, as well as mathematicians,
we have presented many details that would not appear in a
research report and we have excluded proofs. (Most of the re
sults can be proved in a straightforward way.) It is assumed,
of course, that the reader has a working knowledge of the
rudiments of classical calculus.

Jane Grossman
Michael Grossman
Robert Katz

July 14, 1980

 CONTENTS

PRELIMINARIES, 0

CHAPTER 1 CLASSICAL CALCULUS

1.0 Introduction, 1
1.1 Arithmetic Average, 1
1.2 Linear Functions, 2
1.3 Classical Slope, 3
1.4 Classical Gradient, 4
1.5 Classical Derivative, 5
1.6 The Basic Theorem of Classical Calculus, 6
1.7 The Basic Problem of Classical Calculus, 7
1.8 Classical Integral, 8
1.9 The Fundamental Theorems of Classical Calculus, 10

CHAPTER 2 WEIGHTED CLASSICAL CALCULUS

2. Introduction, 11
2. u-Measure, 11
2. u-Average, 12
2. u-Uniform Functions, 16
2. u-Slope, 17
2. u-Gradient, 17
2. u-Derivative, 19
2. The Basic Theorem of u-Calculus, 21
2. The Basic Problem of u-Calculus, 23
2. u-Integral, 24
2.10 The Fundamental Theorems of u-Calculus, 27
2.ll An Unexpected Result, 29

CHAPTER 3 WEIGHTED NON-NEWTONIAN CALCULUS

3.0 Introduction, 30
3.1 u*-Measure, 31
3.2 u*-Average, 32
3.3 u*-Uniform Functions, 35
3.4 u*-Slope, 37
3.5 u*-Gradient, 37
3.6 u*-Derivative, 39
 3.7 The Basic Theorem of u*-Calculus, 41
3.8 The Basic Problem of u*-Calculus, 43
3.9 w*~ Integral, 44
3.10 The Fundamental Theorems of u*-Calculus, 47
3.ll Relationship between u*-Calculus and Weighted
Classical Calculus, 48
3.12 An Alternative Approach to u*-Calculus, 49
3.13 Weighted Geometric Calculus, 49

LIST OF SYMBOLS, 52

INDEX, 5Z
 PRELIMINARIES

The word number means real number. The letter R stands

for the set of all numbers.
If r < s, then the interval [r, s] is the set of all
numbers x such that r g x g s. (Only such intervals are used
here.) The interior of [r, s] consists of all numbers x such
that r < x < s. The classical extent of [r, s] is s - r. A
unit interval is any interval whose classical extent is 1.
An arithmetic partition of [r, s] is any arithmetic
progression whose first term is r and last term is s. An ar
ithmetic partition with exactly n terms is said to be n-fold.
A point is any ordered pair of numbers, each of which
is called a coordinate of the point. A function i& any set of
points, each distinct two of which have distinct first
coordinates.
The domain of a function is the set of all its arguments
(first coordinates) ; the range of a function is the set of all
its values (second coordinates) . A function is said to be on
its domain, to be onto its range, and to be defined at each of
its arguments.
The classical change of a function f on [r, s] is
f (s) - f (r).
If every two distinct points of a function f have dis
tinct second coordinates, then f is one-to-one and its inverse,
f , is the one-to-one function consisting of all points (y, x)
for which (x, y) is a point of f.
A positive function is any function whose values are all
positive; a discrete function is any function that has only a
finite number of arguments.
The function exp is on R and assigns to each number x
the number ex, where e is the base of the natural logarithm
function, In. The function In is the inverse of exp.
"for each successive class of phenomena,
a new calculus or a new geometry, as
the case might be, which might prove
not wholly inadequate to the subtlety
of nature. "

Quoted, without citation,

by H. J. S. Smith; Nature,
Volume 8 (1873), page 450.
 CHAPTER 1

Classical Calculus

1.0 INTRODUCTION

In this chapter we present the basic ideas of classical

calculus in a novel manner that leads naturally to the subse
quent construction of the systems of weighted calculus.
Although the term "classical analysis" is often used,
the term "classical calculus" has appeared rarely, presumably
because only one system of calculus was known prior to the
discovery of the non-Newtonian calculi in 1967.

1.1 ARITHMETIC AVERAGE

The definition of the arithmetic average of a continuous

function on an interval is based on the arithmetic average of n
numbers, which is defined next.

The arithmetic average of n numbers y, , . . . , y is the

number (y, + • • • + y ) / n .

Let f be any continuous function on [r, s] . Then the

arithmetic average of f on [r, s] is denoted by Msf and is de
fined to be the limit of the convergent sequence whose nth term
is the arithmetic average

[f (a1) + ••• + f (an)] / n ,

where a^, . . . , an is the n-fold arithmetic partition of [r, s] .

 The operator M is
Additive: M°(f + g) = M^f + M^g ,
Homogeneous : MS(c • f) = c • Msf , c any constant.

The operator M is characterized by the following three

properties. (This use of the term "characterized" indicates
that no other operator possesses all three properties.)
1. For any interval [r, s] and any constant function k(x) = c
on [r, s] ,
M^k = c .

2. For any interval [r, s] and any functions f and g continu

ous on [r, s] , if f{x) S g(x) on [r, s] , then

3. For any numbers r, s, t such that r < s < t and any func
tion f continuous on [r, t] ,

(s - r) • Mf + (t - s) • Mf = (t - r) • M
-L O L

1.2 LINEAR FUNCTIONS

In classical calculus the linear functions are the stan

dard to which other functions are compared.
A linear function is any function on R expressible in
the form bx + c, where b and c are constants.
Each linear function u is 'uniform' in the following
sense:
 for any intervals [r,, s,] and [r^, s^] with the same
classical extent,
u(s,) - uir^) = u(s2) - u(r2);

that is, u has the same classical change on all inter

vals with the same classical extent.

Notice, in particular, that each linear function has the

same classical change on all unit intervals.
In Section 1.4, we shall exploit the fact that there is
exactly one linear function containing any two given points
with distinct first coordinates.
It can easily be proved that if u is a linear function,
then the ratio
u(s) - u(r)

has the same value on EVERY interval [r, s] . That fact sug
gests the definition of classical slope given next.

1.3 CLASSICAL SLOPE

The classical slope of a linear function u is the common

value of all the ratios
u(s) - u(r)
s - r

where r < s. It is worth noting that the classical slope of u

is equal to its classical change on any unit interval.
Of course, the classical slope of the linear function
bx + c turns out to be b.
1.4 CLASSICAL GRADIENT

The differential branch of classical calculus is rooted

in the concept of the average rate of change, which we prefer
to call the classical gradient. (The term "gradient" is better
suited for our purpose, since it can readily be modified by
appropriate adjectives. Of course, "gradient" is also used in
vector analysis, but that subject does not concern us here.)

Let f be any function defined at r and s, where r < s.

Then the classical gradient of f on [r, s] is denoted by Gsf

and is defined to be the classical slope of the linear func

tion containing the points (r, f(r)) and (s, f(s)). It turns
out that
s f(s) - f(r)
r

If u is a linear function, then G^u equals the classical

slope of u.

The operator G is

Additive: G^(f + g) = G^f + G^g ,

Homogeneous : GS(c • f) = c • Gsf , c any constant.

Finally, when r = s, the expression for the classical

gradient yields the indeterminate form 0/0, which brings us
to the next topic.
1.5 CLASSICAL DERIVATIVE

Let f be any function defined at least on an interval

containing the number a in its interior. If the following lim
it exists, we denote it by [Df] (a) and call it the classical
derivative of f at a:

f(x) - f(a)
lim
x - a

Notice that if a < x, then

f(x) - f(a)

x - a

And if x < a, then

f(x) - f(a)

The classical derivative of f, denoted by Df, is the

function that assigns to each number t the number [Df ] (t) ,
if it exists.
The classical derivative of a linear function has a
constant value equal to its classical slope. Indeed, only lin
ear functions have classical derivatives that are constant on
R. In particular, if u is a function that is constant on R,
then Du is everywhere equal to 0.

The operator D is
Additive: D(f + g) = Df + Dg ,
Homogeneous : D(c • f) = c • Df , c any constant.
1.6 THE BASIC THEOREM OF CLASSICAL CALCULUS

For many years we have been guided by the idea that the
"kernel" of the classical calculus is a well-known result that
we call the Basic Theorem of Classical Calculus. Let us begin
with its discrete analogue, which is a proposition that con
cerns discrete functions and appropriately conveys the spirit
of the theorem.

Discrete Analogue of the Basic Theorem of Classical Calculus

Let h be a discrete function whose arguments a,, ..., a
constitute an arithmetic partition of [r, s] . Then

a
G£ i_ i , r* n h
a i_
h +•••+«
a a
1 n-1 _
; = Grh '•
n-1

that is, the arithmetic average of the n-1 successive

classical gradients of h is equal to the classical gra
dient of h on [r, s] .

The preceding result suggests the following important

theorem.

Basic Theorem of Classical Calculus

If Dh is continuous on [r, s] , then its arithmetic aver
age on [r, s] is equal to the classical gradient of h on
[r, s] ; that is,
h(s) - h(r)
M(Dh) = G
s - r
 In view of the preceding theorem it is appropriate to
say that the arithmetic average fits naturally into the scheme
of classical calculus.

1.7 THE BASIC PROBLEM OF CLASSICAL CALCULUS

Suppose that the value of a function h is known at an

argument r, and suppose that f, the classical derivative
of h, is continuous and known at each number in [r, s] .
Find h(s) .
Solution
By the Basic Theorem of Classical Calculus,

h(s) - h(r)

Solving for h(s), we get

h(s) = h(r) + (s - r)

The number (s - r) • M f that appears in the preceding

solution will arise with sufficient frequency to warrant a spe
cial name, "the classical integral of f on [r, s]," which is
introduced in the next section.

Thus, the Basic Theorem of Classical Calculus, which

involves the arithmetic average, classical derivative, and
classical gradient, provides for the Basic Problem of Classical
Calculus an immediate solution, which, in turn, motivates our
definition of the classical integral in terms of the arithmetic
average.
1.8 CLASSICAL INTEGRAL

Let f be any continuous function on [r, s] . Then the

classical integral of f on [r, s] is denoted by / f and is
defined to be the number
(s - r) • M?f .

We set / f = 0 and f f = — f f .
Jr Js Jr

The operator / is
_-s /-s /.s
Additive : /"(f + g)= f f + f 9 ,
Jr JT Jr

Homogeneous : /f.3 (c • f) = c • /f.3 f , c any constant.

•/— J~

rs
The following result provides a way of expressing / f
•'r
as the limit of a sequence of sums.
rs
If f is continuous on [r, s] , then / f equals the
Jt
limit of the convergent sequence whose nth term is
the sum
jn.f<ai) + ... t VfU^) ,

where a,, ..., a is the n-fold arithmetic partition of

 [r, s] , and j is the common value of

a2 " al' a3~ 32' ••• ' an an-l '

The operator / is characterized by the following three

properties:

1. For any interval [r, s] and any constant function k(x) = c

on [r, s] ,
s
k = c • (s - r) .
r

2. For any interval [r, s] and any functions f and g continu

ous on [r, s] , if f(x) g g (x) on [r, s] , then

-s
g •
r

3. For any numbers r, s, t such that r < s < t and any func

tion f continuous on [r, t] ,

-s
f = / f .
s •'r
10

1.9 THE FUNDAMENTAL THEOREMS OF CLASSICAL CALCULUS

The classical derivative and classical integral are

'inversely' related in the sense indicated by the following
two theorems, the second of which is a simple consequence of
the Basic Theorem of Classical Calculus discussed in Section
1.6.

First Fundamental Theorem of Classical Calculus

If f is continuous on [r, s] , and

g(x) = /f* f , for every number x in [r, s] ,

•'r

then
Dg = f , on [r, s] .

Second Fundamental Theorem of Classical Calculus

If Dh is continuous on [r, s] , then
•s
(Dh) = h(s) - h(r) .
 CHAPTER

Weighted Classical Calculus

2.0 INTRODUCTION

A classical weight function is any function that is con

tinuous and positive on R. There are, of course, infinitely-
many classical weight functions.
For each classical weight function u , we shall develop
the u-calculus, which is a weighted classical calculus in the
sense that its operators are the classical calculus operators
"weighted" by u. It turns out that if u(x) = 1 on R, then the
u-calculus is identical with the classical calculus.

2.1 u-MEASURE

For the remainder of this chapter, u is an arbitrarily

rx
selected classical weight function, and W(X) = I u for each
J0
number x.
Notice that W is a one-to-one function on R and that
DW = w.
The u-measure of an interval [r, s] is denoted by
m[r, s] and is defined to be the positive number
•s
If W ,
Jr

which equals W(s) - V(r).

Notice that if u(x) =1 on R, then the u-measure of each
interval [r, s] is equal to its classical extent s - r.

ll
12

A w-unit interval is any interval whose u-measure is 1,

There may or may not exist u-unit intervals, depending on u.

2.2 u-AVERAGE

The definition of the u-average of a continuous function

on an interval is based on the weighted arithmetic average of n
numbers, which is defined next.

Let y, , . . . , y be any n numbers. Then for any n posi

tive numbers v, , . . . , v , the weighted arithmetic average of
y y is

v, + • • • + v
1 n

Let f be any continuous function on [r, s] . Then the

u-average of f on [r, s] is denoted by t^^f and is defined to be
the limit of the convergent sequence whose nth term is the
weighted arithmetic average

u(a1).f(a1) + •-• + u(an).f(an)

+ • • • + w (an)

where a, , . . . , a is the n-fold arithmetic partition of [r, s]

 13

It turns out, of course, that ^sf is the well-known

weighted arithmetic average of f on [r, s] :
s
L r
( w • f)

Also, 14 f equals
-s
(f
r
s
DW

which is the arithmetic average of f RELATIVE to W on [r, s]

If w is constant on R, then Msf reduces to the
(unweighted) arithmetic average Msf, which equals

s
f

s - r

The operator M is
Additive: ^S (f + g) = ^Sf + MSg ,

Homogeneous : Ms (c • f ) = c • Msf , c any constant.

14

The operator ^ is characterized by the following three

properties:

1. For any interval [r, s] and any constant function k(x) = c

on [r, s] ,

2. For any interval [r, s] and any functions f and g continu

ous on [r, s] , if f(x) 5; g(x) on [r, s] , then

3. For any numbers r, s, t such that r < s < t and any func
tion f continuous on [r, t] ,

m[r, s] .M^f + m[s, t] -M^f = m[r, t] -M^f .

— JL —S —i

A u-partition of an interval [r, s] is any finite se

quence of numbers a,, ..., a such that

r = a.^ < a2 < • • • < an = s

and
m[a,, a~] = m[a-, a.,] = ••• = m[a ,, a ] .

A u-partition with exactly n terms is said to be n-fold.

 15

For each continuous function f on [r, s] , we defined M f

to be the limit of the convergent sequence whose nth term is
the WEIGHTED arithmetic average

u(an)-f(an)

+ • • • + w (an)

where a, , . . . , a is the n-fold arithmetic partition of [r, s]

Hence we were surprised to discover that M f also equals the
limit of the convergent sequence whose nth term is the
(UNWEIGHTED) arithmetic average

f(an)

where a, , . . . , a is the n-fold u-partition of [r, s] .

We close this section with a simple result worth noting.

Let f be any function continuous on [r, s] , and let c

be any positive constant. Then cu is a classical
weight function and on [r, s] , the cu-average of f is
equal to the u-average of f.
16

2.3 u-UNIFORM FUNCTIONS

In classical calculus, the linear functions are the

standard to which other functions are compared; in u-calculus
the standard is provided by the u-uniform functions, which will
be defined shortly.
rx
Recall that we defined W (x) = I w , for each number x.
J0
Now consider any function u such that
u(x) = b-V(x) + c , on R,
where b and c are constants.
It turns out that for any intervals [r,, s-^] and
[r2/ s2] with the same u-measure,
u(s1) - u(r^) = u(s2) - u(r2) ;
that is, u has the same classical change on all intervals with
the same u-measure. Therefore each function on R expressible
in the form b'V(x) + c, and only such a function, is called a
w-uniform function.
Notice, in particular, that each u-uniform function has
the same classical change on all u-unit intervals, if any.
Clearly every function that is constant on R is w-
uniform (take b = 0) .
Moreover, if w is constant, then the u-uniform functions
are identical with the linear functions.
In Section 2.5, we shall exploit the fact that there is
exactly one u-uniform function containing any two given points
with distinct first coordinates.
 17

It can easily be proved that if u is a u-uniform func

tion, then the ratio
u(s) - u (r)
m[r, s]
has the same value on EVERY interval [r, s] . That fact sug
gests the definition of u-slope given next.

2.4 u-SLOPE

The u>-slope of a u-uniform function u is the common

value of all the ratios
u(s) - u(r)
m[r, s]

It is worth noting that the u-slope of u is equal to its clas

sical change on any u-unit interval, if any.
The u-slope of the u-uniform function b-V(x) + c turns
out to be b.

2.5 u-GRADIENT

The differential branch of u-calculus is rooted in the

concept of the u-gradient, which is defined next.
Let f be any function defined at r and s, where r < s.
Then the u-gradient of f on [r, s] is denoted by Q^f and is
defined to be the u-slope of the u-uniform function containing
the points (r, f(rO) and (s, f(s)). It turns out that
f(s) - f(r) f(s) - f(r)
irf = = = •
m[r, s]
18

If u is a u-uniform function, then g u equals the u-

slope of u.
Since
•8
u = V(s) - W(r) ,
*^3T

we have
f(s) - f(r)
W(s) - W(r)
which is the classical gradient of f RELATIVE to W on [r, s]

If u(x) = 1 on R, then
-s
u = s - r ,
r
and so
irf = Grf '

The operator g is
Additive; g^(f + g) = g^f + g^g ,
Homogeneous : g S (c*f) =c-Gf,
S
c any constant.

Finally, when r = s, the expression for the u-gradient

yields the indeterminate form 0/0, which brings us to the next
topic.
 19

2.6 u-DERIVATIVE

Let f be any function defined at least on an interval

containing the number a in its interior. If the following
limit exists, we denote it by [gf] (a) and call it the
w-derivative of f at a:

f(x) - f(a)

w
a
Notice that if a < x, then

f(x) - f(a)

a
And if x < a, then

f(x) - f(a)

Jarx w
It can be proved that [gf] (a) and [Df] (a) coexist; that
is, if either exists then so does the other. Moreover, if they
do exist, then

[Df ] (a) [Df ] (a)

[Df ] (a) =
u(a) [Df/] (a)

If u(x) =1 on R, then [gf ] (a) = [Df ] (a) .

The u-derivative of f, denoted by gf, is the function

that assigns to each number t the number [Df] (t) , if it exists.
20

Since I u = V(x) - V(t) , we have

f(x) - f(t)
[gf ] (t) = lim .
x^fc W(x) - (/(t)

Hence gf equals the classical derivative of f RELATIVE to W.

The u-derivative of a u-uniform function has a constant
value equal to its u-slope. Indeed, only u-uniform functions
have u-derivatives that are constant on R. In particular, if
u is a function that is constant on R, then gu is everywhere
equal to 0.
The operator g is
Additive: g(f + g) = gf + Dg ,
Homogeneous : D(c • f) = c • Df , c any constant.

In the classical calculus, Dh = h if h (x) = exp x. In

the u-calculus, gh = h if h(x) = exp[V(x)].
It is worth noting that the u-derivative of a is equal
to the logarithmic derivative of w, that is,
gu = [Du]/u = D(ln u) .
We conclude this section with the Mean Value Theorem of
u-Calculus. (Also provable in the u-calculus are analogues of
other classical theorems such as L' Hospital's Rule and Taylor's
Theorem. )

Mean Value Theorem of u-Calculus

If a function f is continuous on [r, s] and [gf] (x)
exists for each number x between r and s, then between
r and s there is a number c such that

f(s) - f(r)
[Df ] (c) = g^f = .
m[r, s]
 21

NOTE

The following restatement of the preceding theorem is an

immediate consequence of Cauchy's well-known Generalized Mean
Value Theorem of Classical Calculus:
If f is continuous on [r, s] and its classical deriva
tive exists at every number between r and s, then be
tween r and s there is a number c such that

[Df] (c) f (s) - f (r)

[DW] (c) (/(s) - W(r)

2.7 THE BASIC THEOREM OF u-CALCULUS

Just as there is a discrete analogue of the Basic

Theorem of Classical Calculus, so there is a

Discrete Analogue of the Basic Theorem of u-Calculus

Let h be a discrete function whose arguments a,, ..., a

constitute an arithmetic partition of [r, s] . For each
integer i from 1 to n - 1 , let

Then

v, (G h) + ••• + v__n ( G r h)

vn-l

that is, a certain weighted arithmetic average of the

n - 1 successive u-gradients of h is equal to the in
gredient of h on [r, s] .
22

The preceding result suggests the following important

theorem.

Basic Theorem of u-Calculus

If Dh is continuous on [r, s] , then its u-average on

[r, s] is equal to the u-gradient of h on [r, s] ;
that is,

h(s) - h(r)
m[r, s]

In view of the preceding theorem it is appropriate to

say that the u-average fits naturally into the scheme of u-
calculus.

NOTE
Here is another discrete analogue of the Basic Theorem
of u-Calculus:
Let h be a discrete function whose arguments
a,, ..., a constitute a u-partition of [r, s]. Then

Ga2h + ... + Gan h

- al - an-l

n - 1

that is, the (unweighted) arithmetic average of the

n - 1 successive u-gradients of h is equal to the u-
gradient of h on [r, s] .
 23

2.8 THE BASIC PROBLEM OF u-CALCULUS

Suppose that the value of a function h is known at an

argument r, and suppose that f, the u-derivative of h,
is continuous and known at each number in [r, s] .
Find h(s) .
Solution
By the Basic Theorem of u-Calculus,
h(s) - h(r)
M"f =
m[r, s]
Solving for h(s) , we get
h(s) = h (r) + m[r, s] -Msf .

The number m[r, s]-MSf that appears in the preceding

solution will arise with sufficient frequency to warrant a
special name, "the u-integral of f on [r, s]," which is intro
duced in the next section.

Thus, the Basic Theorem of u-Calculus, which involves

the u-average, u-derivative , and u-gradient, provides for the
Basic Problem of u-Calculus an immediate solution, which, in
turn, motivates our definition of the u-integral in terms of
the u-average.
24

2.9 u-INTEGRAL

Let f be any continuous function on [r, s] . Then the

rs
^-integral of f on [r, s] is denoted by I f and is defined
J•*-
to be the number

m[r, s] • M^f .

We set / f = 0 .

It turns out that

-s /.s /-s
If - / (f • «0 - / (» • f J
^r •'r •'r

which is a weighted classical integral, and which in certain

contexts is called an inner product.

If u(x) = 1 on R, then

The operator / is
rs
Additive: / (f + g) =

Homogeneous : / (c • f) = c • / f ,
£x ?r
c any constant.
 25

The following results provide two ways of expressing

fs
I £ as the limit of a sequence of sums.
= r
rs
1. If f is continuous on [r, s] , then / f equals the limit
= r
of the convergent sequence whose nth term is the sum

Jn-U(a1)-f(a1) + ••• + V^n-l'-^n-l* '

where a., ..., a is the n-fold arithmetic partition of

[r, s] , and j is the common value of

a2 ~ al' a3 ~ a2' ••' ' an " an-l '

fS
2. If f is continuous on [r, s] , then / f equals the limit
= r
of the convergent sequence whose nth term is the sum

k -f (a,) + ••• + k -f (a .) ,

where a^, . . . , an is the n-fold u-partition of [r, s] , and

k n is the common value of

- V(ai) for i = 1 , . . . , n - 1 , it is

not surprising that

rs
I f is the Stieltjes integral of f
=r
RELATIVE to V on [r, s].
26

The following two results are also worth noting.

If f is continuous on [r, s] and c is any positive
constant, then on [r, s], the cu-integral of f is
equal to the product of c and the u-integral of f.
If f is continuous on [r, s] and w, and w2 are classi
cal weight functions with sum w on R, then on [r, s] ,
the sum of the w, and w_-integrals of f is equal to
the u-integral of f.

The operator / is characterized by the following

three properties:

1. For any interval [r, s] and any constant function k(x) = c

on [r, s] ,
rs
I k = c • m [r , s].
=r

2. For any interval [r, s] and any functions f and g continu

ous on [r, s] , if f (x) g g (x) on [r, s], then
rs rs
J f = J g •
= r Jfr

3. For any numbers r, s, t such that r < s < t and any func
tion f continuous on [r, t] ,

/"'*/*'-
= r =rs /*
^r
 27

We conclude this section with another mean value theorem

of u-calculus:
If f is continuous on [r, s] , then between r and s
there is a number c such that

l£f = f(c) ;

that is, the u-average of a function continuous on

[r, s] is assumed at some argument between r and s.

NOTE

The following restatement of the preceding mean value

theorem is an immediate consequence of the well-known Second
Mean Value Theorem for Classical Integrals:
If f is continuous on [r, s] , then between r and s
there is a number c such that
•s (.s
u) = f (c) / w .
*/ -V-

2.10 THE FUNDAMENTAL THEOREMS OF u-CALCULUS

The u-derivative and u-integral are 'inversely1 related

in the sense indicated by the following two theorems, the sec
ond of which is a simple consequence of the Basic Theorem of
u-Calculus discussed in Section 2.7.
28

First Fundamental Theorem of w-Calculus

If f is continuous on [r, s] , and

g(x) = / f , for every number x in [r, s] ,

= r
then
gg = f , on [r, s] .

Second Fundamental Theorem of w-Calculus

If Dh is continuous on [r, s] , then
•s
rs
I (gh) = h(s) - h(r) .
JT?

Just as the Second Fundamental Theorem of Classical Cal

culus is useful for evaluating classical integrals, the Second
Fundamental Theorem of u-Calculus is useful for evaluating u-
integrals. As an example, let u(x) = x2 + 1, f(x) = 4x, and
h(x) = x4 + 2x2 + 2, for every number x. Then f = gh, and so

1 /-I
= h(l) - h(0) =5-2=3.
Jn
 29

2.ll AN UNEXPECTED RESULT

In this section we assume that w is not identically 1 on

R. (If u is identically 1 on R, then the u-calculus is the
classical calculus, which does not concern us in this section.)
We also assume that the reader is familiar with Chapters 5 and
6 of Non-Newtonian Calculus.
rx
Recall that we defined W(x] I w, for each number x.
0 -1
Now assume that the range of W is R, and let a = W , which is

a one-to-one function whose domain and range equal R. Then a

generates a-arithmetic, whose realm is R.

In Non-Newtonian Calculus it is shown that the ordered

pair of arithmetics (a-arithmetic, classical arithmetic) deter
mines an (UNWEIGHTED) non-Newtonian calculus. Surprisingly,
that system is identical with the u-calculus, which is a
WEIGHTED classical calculus.

****************************************************************

Since weight functions are usually assumed to be positive,

we developed the u-calculus only for the case where w is a con
tinuous positive function on R. However, it is possible to de
velop the u-calculus for the case where u is a continuous nega-
tive function on R.
 CHAPTER 3

Weighted Non-Newtonian Calculus

3.0 INTRODUCTION

In this chapter, we shall use the notation, terminology,

and results set forth in Non-Newtonian Calculus.
For the remainder of the chapter, a and B are arbitrari
ly selected generators and * is the ordered pair of arithmetics
(a-arithmetic, B-arithmetic) . The following notations will be
used.

a-arithmetic B-arithmetic

Realm A B
Addition + +

Subtraction - -
Multiplication * x

Division / /or

Order < <

We denote the isomorphism from a-arithmetic to B-arithmetic

by i .

A *-weight function is any function that is *-continuous

and B-positive on A. (Of course, a "B-positive function" is a
function whose values are all B-positive numbers.) There are
infinitely-many *-weight functions.

30
 31

For each *-weight function w, we shall develop the u*-

calculus, which is a weighted *-calculus in the sense that its
operators are the *-calculus operators "weighted" by w. It
turns out that if u(x) = 1 on A, then the u*-calculus is iden
tical with the *-calculus. Also, in the special case for which
a(x) = x = B(x) on R, the u*-calculus turns out to be a
weighted classical calculus.
The operators of each u*-calculus are applied only to
functions with arguments in A and values in B. Accordingly,
unless indicated or implied otherwise, all functions considered
in this chapter are assumed to be of that character.
We close this section with a notational convention,
which will be used shortly.
For each *-continuous function f on an a-interval
[r, sj,
*r * s
f f
Js
denotes 6 - I
JT
f

3.1 u*-MEASURE

For the remainder of this chapter, w is an arbitrarily

selected *-weight function, and
*x
V(x) = f w ,
JQ

for each number x in A.

Then W is a one-to-one function on A, and DW = u .
32

The wA-measure of an a-interval [r, s] is denoted by

*.
m[r, s] and is defined to be the B-positive number

r
I
JT

which equals V(s) - W(r) .

A u*-unit interval is any a-interval whose u*-measure is

1. There may or may not exist u*-unit intervals, depending on
w.

3.2 u*-AVERAGE

Our definition of the u*-average of a *-continuous func

tion on an a-interval is based on the weighted B-average of n
numbers, which we define next.

Let y, , . . . , y be any n numbers in B. Then for any n

B-positive numbers v, , . . . , v , the weighted S~average of
y, , . . . , y is the following number in B:

Let f be any function *-continuous on [r, sj . Then the

u*-average of f on [r, s] is denoted by 5}*s f and is defined to
be the B-limit of the B-convergent sequence whose nth term is
the weighted B-average

] + ••• + [u(an) x f(an)]

 33

where a,, ..., a is the n-fold a-partition of [r, s]

*s
It turns out that ^ f equals

-
If (w x f)
Jr

Since
*
.. *
(f x D
/^
.*.s.
*e
*
X DW

we may call M f "the *-average of f RELATIVE to W on [r, s]."

*s
If w is constant on A, then M
"' f reduces to the
(unweighted) *-average M*s f .

In certain special cases the u*-average is well-known.

*s f is the weighted
For example, if a(x) = x = B(x) on R, then J}
arithmetic average of f on [r, s] (see Section 2.2). If a(x)
*s
= x and B(x) = exp x, on R, then M f is the weighted geometric
average of f on [r, s] . Also, if ot(x) = x on R,
''P for x > 0
! (x) = JO
{0 for x = 0 , p ^ 0 ,
'
- (- x)p for x < 0

and f is positive, then M*s f is the weighted pth-power average

of f on [r, s] , which, if p = -1, is the weighted harmonic
average of f on [r, s] .
34

*
The operator M is
B-Additive: g*{f + g) = MSf + g ,
*s •• •• *s
B-Homogeneous: M (c x f) = c * M f ,

c any constant in B.

The operator M is characterized by the following three

properties:

1. For any a-interval ['r, sj and any constant function k(x) =

c on fr, s] ,

2. For any a-interval [r, sj and any functions f and g *-

continuous on [r, sj , if f(x) g g(x) on [r, s] , then
"ft c .• "ft c
M f S M g .

3. For any numbers r, s, t in A such that r < s < t and any

function f *-continuous on [r, t j ,

(m[r, sj x 0 f) + (m[s, tj x M f) =m[r, t] x y f .

A u*-partition of an a-interval [r, si is any finite

sequence of numbers a.^, . . . , an in A such that

r = a, < a~ < • • • < a =s

and

A u*-partition with exactly n terms is said to be

n-fold.
 35

For each function f *-continuous on [r, s] , we defined

Msf to be
be the
the B-limit
B-limit of
of the
the 00-convergent sequence whose nth
term is the WEIGHTED B-average

[u(an) x f(an)]

where a, , . . . , a is the n-fold a-partition of [r, s] . Never-

theless, it turns out that M*s f also equals the 8-limit of the
B-convergent sequence whose nth term is the (UNWEIGHTED) B~
average
f <ai) + ••• + f (an)

B(n)

where a. , . . . , a is the n-fold u*-partition of [r, s] .

We close this section with a simple result worth noting.

Let f be any function *-continuous on [r, s] , let c be
any B~positive constant, and let v = c x u. Then v is
a *-weight function and on [r, s], the v*-average of f
is equal to the u*-average of f.

3.3 u*-UNIFORM FUNCTIONS

In the u*-calculus, the u*-uniform functions, which will

be defined shortly, are the standard to which other functions
are compared.
ix
Recall that we defined W(x) = I u for each number x
J&
in A.
36

Now consider any function u such that

u(x)= [b * V ( x )] + c , on A,

where b and c are constants in B.

It turns out that for any a-intervals [r,, s,j and
[^2> S2J with the same u*-measure,

ufs-j^) - u(r^) = u(s2) - u(r2) ;

that is, u has the same B-change on all ex-intervals with the
same u*-measure. Therefore each function on A expressible in
the form [b * V(x}] + c , and only such a function, is called a
k;*-uniform function.
Notice, in particular, that each u*-uniform function has
the same B-change on all u*-unit intervals, if any.
Clearly every function that is constant on A is w*-
uniform (take b = 0) .
Moreover, if w is constant, then the u*-uniform func
tions are identical with the *-uniform functions.
In Section 3.5, we shall exploit the fact that there is
exactly one u*-uniform function containing any two given *-
points with distinct first coordinates.
It can easily be proved that if u is a u*-uniform func
tion, then the B-ratio
u(s) - u(r)
* t
m[r, s]4

has the same value on EVERY a-interval [r, s] . That fact sug
gests the definition of u*-slope given next.
 37

3.4 u*-SLOPE

The u*-slope of a o*-uniform function u is the common

value of all the B-ratios

u(s) - u(r)

m['r, s]

It is worth noting that the u*-slope of u is equal to its 6~

change on any u*-unit interval, if any.
The u*-slope of the u*-uniform function [b x V(x)] + c
turns out to be b.

3.5 u*-GRADIENT

The differential branch of u*-calculus is rooted in the

concept of u*-gradient, which is defined next.

Let f be any function defined at r and s, where r < s .

*s
Then the ij*-gradient of f on [r, s] is denoted by g f and is
defined to be the u*-slope of the u*-uniform function contain
ing the *-points (r, f(r)) and (s, f(s}). It turns out that

*_ f(s) - f(r) f(s) - f(r)

rs .
m[r, s]
JrI 1}
If u is a u*-uniform function, then g u equals the u*
slope of u.
38

rs
Since / w = W(s) - W(r) ,
Jt
we have
*_ f(s) - f(r)
g'f =

*s
and so we may call grf "the *-gradient of f RELATIVE to W on

t'r, s] . "

If w(x) = 1 on A, then
rs
I w = i (s) - i (r) ,
•'r
*s • t
and so Q^f equals the *-gradient of f on [r, s] .

*
The operator Q is
B-Additive; |^(f + g) = g°f + Q^g ,
*s •• •• *s
B-Homogeneous; g (c*f) =cxgf,
c any constant in B.

Finally, when r = s, the expression for the u*-gradient

yields the indeterminate form 0 / 0 , which brings us to the
next topic.
 39

3.6 u*-DERIVATIVE

Let f be any function defined at least on an a-interval

containing the number a in its a-interior. If the following
*-limit exists, we denote it by [gf] (a) and call it the
ij*-derivative of f at a:

f(x) - f(a)
*-lim
x-t-a

Notice that if a < x, then

f(x) - f(a) t
sx
'a
And if x < a, then

f(x) - f(a) t
*..
= G*f
X
.

If it exists, the number [gf] (a) is in B.

* *
It can be proved that [gf] (a) and [Df ] (a) coexist.
Moreover, if they do exist, then

A [Df ] (a) [Df ] (a)

[Df] (a) =
u(a) [DW] (a)

If u(x) =1 on A, then [Df] (a) = [Df] (a) .

40

*
The u*-derivative of
— - f, denoted by Df,
.- is the function
^
that assigns to each number t in A the number [Df ] (t) , if it
exists.
Since
/:
/ u = V(x) - V(t), we have

[Df ] (t) = *-lim

f(x) - f(t>

f/(x) - V(t)

Hence gf may be called "the *-derivative of f RELATIVE to W."

The u*-derivative of a u*-uniform function has a con

stant value equal to its u*-slope. Indeed, only u*-uniform
functions have u*-derivatives that are constant on A. In par-
*
ticular, if u is a function that is constant on A, then gu = 0
on A.
*
The operator g is
B-Additive: g(f + g) = gf + gg ,
* .. *
6-Homogeneous : D(c x f) = c x gf ,
c any constant in B.

It is worth noting that if a(x) = exp x and B (x) = x, on

R, then the u*-derivative of u is equal to the so-called elas
ticity of w. (The elasticity at an argument a of a function f
with positive arguments and values is defined to be

[f(x) - f(a)]/f(a)
lim ,
(x - a)/a

which, if it exists, equals a • [Df ] (a) /f (a) . Elasticity is

often used in economics.)
 41

We conclude this section with the Mean Value Theorem of

u*-Calculus. (Also provable in the u*-calculus are analogues
of other classical theorems such as L' Hospital's Rule and Tay
lor1 s Theorem. )

Mean Value Theorem of u*-Calculus

*
If a function f is *-continuous on [r, s] and [Df ] (x)
exists for each number x such that r < x < s, then there
exists a number c for which r < c < s and

[Df ] (c) = *sSf = f(s) - f(r) .

* r
m[r, s]i

3.7 THE BASIC THEOREM OF u*-CALCULUS

We begin with a discrete analogue. (Another discrete

analogue is provided in the Note at the end of this section.)

Discrete Analogue of the Basic Theorem of u*-Calculus

Let h be a discrete function whose arguments a, , . . . , a
constitute an a-partition of [r, s] . For each integer
i from 1 to n - 1, let

Then

* 37
<V, x 5/ hi +•..+( v , x G* a n h)
-L -* a _ IT"' -L — a
1 n-l

vn-l
42

that is, a certain weighted -g-average of the n - 1 suc

cessive u*-gradients of h is equal to the u*-gradient of
h on [r , s] .

The preceding result suggests the following important

theorem.

Basic Theorem of u*-Calculus

*
If Dh is *-continuous on [r, s] , then its u*-average on
[r, s] is equal to the u*-gradient of h on [r, s] ;
that is,

*,. * *„ h(s) - h(r)

mfr, s]

In view of the preceding theorem it is appropriate to

say that the u*-average fits naturally into the scheme of w*-
calculus.

NOTE

Here is another discrete analogue of the Basic Theorem

of u*-Calculus:
Let h be a discrete function whose arguments
a,, ..., a constitute a u*-partition of [r, si . Then

*a, .. .. *an
G h + ••• + G h
=a =3

B(n - 1)

that is, the (unweighted) B-average of the n - 1 suc

cessive u*-gradients of h is equal to the u*-gradient
of h on [r, s] .
 43

3.8 THE BASIC PROBLEM OF u*-CALCULUS

Suppose that the value of a function h is known at an

argument r, and suppose that f, the u*-derivative of h,
is *-continuous and known at each number in [r, sj .
Find h(s) .
Solution
By the Basic Theorem of u*-Calculus,

h(s) - h(r)
* f
m[r, s]\

Solving for h(s), we get

h(s) = h(r) + (m[r, sj x

*• • •• s
The number m[r, s] x M f that appears in the preceding
solution will arise with sufficient frequency to warrant a spe
cial name, "the u*-integral of f on [r, sj," which is intro
duced in the next section.

Thus, the Basic Theorem of u*-Calculus, which involves

the u*-average, u*-derivative, and u*-gradient, provides for
the Basic Problem of u*-Calculus an immediate solution, which,
in turn, motivates our definition of the u*-integral in terms
of the u*-average.
44

3.9 u*-INTEGRAL

Let f be any function *-continuous on [r, sj . Then the

m
u*-integral of f on [r, s] is denoted by f/ f and is defined
zx
to be the following number in B:

m[r, s] x £* f .
/*r
VJe set / f = 0 .
**

It turns out that

/•<-/ (u x f) ,
^ 3T r

which might appropriately be called "a weighted *-integral."

Since
rs rs •• *
/ f = / (f x DV) ,
gx Jr

rs
we may call / f "the *-integral of f RELATIVE to W on
Jr
(r, s]."

If u(x) =1 on A, then

*S *E
L f =/r
 45
*
The operator / is

*s *s 4s
B-Additive: / (f + g) = / f + / g,
Jr JTC sr

B-Homogeneous : / (c x f) = c x / f,
J J- J •£

c any constant in B.

The following results provide two ways of expressing

f as the B-limit of a sequence of B-sums.

1. If f is *-continuous on [r, sj , then / f equals the 8"

limit of the B-convergent sequence whose nth term is the

B-sum

where a,, ..., a is the n-fold a-partition of fr, s], and

j is the common value of

2. If f is *-continuous on [r, s], then rs

/
f equals the 8-
£ r
limit of the B-convergent sequence whose nth term is the

B-sum
[k x i(a ) ] + ••• + [k x f (a .)] ,
n I n n-1

where a^, . . . , afi is the n-fold u*-partition of [r, s] , and

k_ is the common value of

46

The following two results are also worth noting.

If f is *-continuous on [r, sj , c is any B-positive

constant, and v = c x u, then on [r, sj , the v*-
integral of f is equal to the B-product of c and the
u*-integral of f.

If f is *-continuous on [r, sj and u and v are it-

weight functions with B-sum w on A, then on [r, sj ,
the B-sum of the u* and v*-integrals of f is equal to
the u*-integral of f.

The operator / is characterized by the following three

properties :
1. For any a-interval fr, sj and any constant function k(x) =
c on f r , s j ,
?s
k = c x m[r , s] .

2. For any a-interval [r, sj and any functions f and g *-

continuous on [r, sj, if f(x) < g(x) on [r, sj, then

3. For any numbers r, s, t in A such that r < s < t and any

function f *-continuous on [r, t] ,
 47

We conclude this section with another mean value theorem

of u*-calculus:
If f is *-continuous on [r, s] , then there is a number
c in A such that r < c < s and

M^f = f(c) ;

that is, the u*-average of a function *-continuous on

[r, s] is assumed at some argument.

3.10 THE FUNDAMENTAL THEOREMS OF u*-CALCULUS

The u*-derivative and u*-integral are 'inversely' re

lated in the sense indicated by the following two theorems, the
second of which is a simple consequence of the Basic Theorem of
u*-Calculus discussed in Section 3.7.

First Fundamental Theorem of w*-Calculus

If f is *-continuous on [r, sj , and

g (x) = r*
/ f , for every number x in [r, s] ,
Jr
then
* . .
Qg = f , on [r, s] .

Second Fundamental Theorem of w*-Calculus

*
If Dh is *-continuous on [r, s] , then
*s
f (gh) = h(s) - h(r) .
48

3.ll RELATIONSHIP BETWEEN u*-CALCULUS AND WEIGHTED CLASSICAL

CALCULUS

For the remainder of this section, let w be the classi

cal weight function such that u(t) = 6 (u(ot(t))) for each
number t. We shall indicate the uniform relationships between
the corresponding operators of u*-calculus and u-calculus.
For each number x in A, let x = a (x) . Let f be a
function with arguments in A and values in B, and set f (t) =

We have the following uniform relationships:

(1) G^f = 6 | Qjf } ,

(2) [Df] (a) = 6 | [gf] (a)

(3) :« - •{&}
(4)
f /
=r r
f =

(In (2), the derivatives [Df ] (a) and [gf ] (a) coexist; in (3)
and (4), we assume f is *-continuous on ['r, sj.)

It is also worth noting that f is a u*-uniform function

if and only if f is a u-uniform function.
The foregoing results suggest that for each theorem in
u-calculus, there is a corresponding theorem in u*-calculus,
and conversely.
 49

3.12 AN ALTERNATIVE APPROACH TO u*-CALCULUS

*x
First recall that we defined F/(x) = I w , for each
J&
number x in A. Now assume that the range of W is B, and let
a = v~ o 6, which is a one-to-one function whose domain is R
and range is A. Then 5 generates 5-arithmetic, whose realm
is A.
Let * be the ordered pair of arithmetics
(o-arithmetic, B-arithmetic) . It can be proved that the opera
tors of *-calculus are identical with the corresponding opera
tors of u*-calculus and that the *-uniform functions are iden
tical with the u*-uniform functions. Indeed, *-calculus is
identical with y*-calculus . Thus, the u*-calculus, which is a
WEIGHTED *-calculus, can be reformulated as an UNWEIGHTED
calculus .
Accordingly, in Non- Newtonian Calculus all the discus
sions, developments, results, interpretations, and heuristic
principles that pertain to the *-calculus are pertinent to the
u*-calculus.

3.13 WEIGHTED GEOMETRIC CALCULUS

In this final section we shall discuss a certain system

of weighted calculus that has many attractive features and may
prove to be one of the most important of all the systems of
calculus.
Henceforth, a(x) = x and 3(x) = exp x, on R, and * is
the ordered pair of arithmetics (a-arithmetic, B-arithmetic) .
Let u be an arbitrarily selected *-weight function.
The u*-calculus is a weighted geometric calculus in the
sense that its operators are the geometric calculus operators
"weighted" by u.
50

The u*-measure of an interval [r, s] equals

exp
i/; In

The u*-average of a continuous positive function f on

[r, s] equals

Ir (In w • In f)
J
exp
In u

which is the well-known weighted geometric average of f on

[r, s].

The u*-uniform functions are expressible in the form

exp In u) + c

where b and c are constants and x is unrestricted in R. The

u*-slope of the preceding function equals exp b .

The u*-gradient of a positive function f on [r, s]

equals

f (s) 1/t

f (r)

where t In u .
 51

The u*-derivative of a positive function f equals

D(ln f)
exp
In u

which is constant on R if and only if f is a u*-uniform

function.

The u*-integral of a continuous positive function f on

[r, s] equals

exp | / (In w - In f)

All the operators of the u*-calculus are multiplicative

and involutional; for example,
* * *
g(f • g) = gf • gg and

* c)
g(f * c ,
= (Df) c any constant .

If u(x) = e on R, then the operators of u*-calculus are

identical with the corresponding operators of the geometric
calculus, the u*-uniform functions are identical with the
geometrically-uniform functions, and> indeed, u*-calculus is
identical with geometric calculus.

In this chapter we developed the u*-calculus only for

the case where w is a *-continuous B-positive function on A.
However, it is possible to develop the u*-calculus for the
case where w is a *-continuous B-negative function on A.
 LIST OF SYMBOLS

Symbol Pag

D 5
D 19
*
D 39, 40
f1 0
6 4

g 17
*
G 37
m 11
*
m 32
H 1

g 12
M 32
R 0
w, V ll, 31
a, B, i 30
[r, s] 0
e, exp, In 0
+ , -, x, /', < 30
, < 30
* 30

8
/

/ 24

/ 44

52
 INDEX

Additive, 2 Defined, 0
Arguments , 0 Discrete analogue, 6
Arithmetic average,! Discrete f unction, 0
relative, 1 3 Domain, 0
weighted ,1 2- 1 3, 33
Arithmetic partition, 0 Elasticity, 40
Exp,0
B-Additive,3<2
6-Homogeneous , 34 Function, 0
Fundamental Theorems :
Basic Problem: Classical Calculus, 10
Classical Calculus,? u-Calculus, 28
u-Calculus, 23 u*-Calculus, 4 7
u*-Calculus ,43
Basic Theorem: Homogeneous, 2
Classical Calculus, 6
u-Calculus,2J-22 Inner product, 24
Interior ,0
Interval, 0
Characterized , 2 Inverse,!?
Classical calculus, 1-10
Classical change, 0 Linear function, 2
Classical derivative, 5 Ln,0
relative, 20 Logarithmic derivative, 20
Classical extent, 0
Classical gradient, 4 Mean Value Theorem:
relative, 18 Cauchy's,22
Classical integral, 5 Classical Integrals, 27
relative , 25 u-Calculus,20, 27
weighted, 24 u*-CaIcuIus,41, 47
Classical slope, 3
Classical weight function, 11 n-Fold, 0,14,34
Coexist, 1 9 Number, 0
Coordinate , 0
53
54

On,0 u-Uniforra function, 16

One-to-one , 0 u-Unit interval, 12
Onto , 0
u*-Average, 32
Point, 0 u*-Calculus, 31
Positive function, 0 u*-Derivative, 33-40
u*-Gradient, 37
Range ,0 w*- Integral, 44
u*-Measure, 32
*-Average: uar-Partition, 34
relative, 33 w*-Slope,37
*-Derivative: u*-Uniform function, 36
relative, 40 u*-Unit interval, 32
*-Gradient :
relative, 38 Weighted arithmetic average, 1 2,
*-Integral : 13,33
relative , 44 Weighted B-average,32
weighted, 44 Weighted classical calculus, 1 1-29
*-Weight function, 30 Weighted geometric average, 33., 50
Weighted geometric calculus, 49-51
Stieltjes integral, 25 Weighted harmonic average, 33
Weighted non-Newtonian calculus,
Unit interval, 0 30-51
Weighted pth-power average, 33
Values, 0

u-Average,I2
u-Calculus, 11
u-Derivative,79
u-Gradient, 1 7
u-Integral, 24
u-Measure, 1 1
u-Partition, 14
u-Slope, 1 7
 In the classical calculus the line is used as a stan
dard against which other curves are compared. In the
following remarks, Roger Joseph Boscovich (1711 - 1787)
may have been the only person to anticipate a fundamen
tal idea in non-Newtonian calculus: a nonlinear curve
may be used as a standard against which other curves
(including lines) can be compared.

'But if some mind very different from ours were to look upon some
property of some curved line as we do on the evenness of a straight
line, he would not recognize as such the evenness of a straight line;
nor would he arrange the elements of his geometry according to that
very different system, and would investigate quite other relationships
as I have suggested in my notes.
'We fashion our geometry on the properties of a straight line
because that seems to us to be the simplest of all. But really all lines
that are continuous and of a uniform nature are just as simple as
one another. Another kind of mind which might form an equally
clear mental perception of some property of any one of these curves,
as we do of the congruence of a straight line, might believe these
curves to be the simplest of all, and from that property of these
curves build up the elements of a very different geometry, referring
all other curves to that one, just as we compare them to a straight
line. Indeed, these minds, if they noticed and formed an extremely
clear perception of some property of, say, the parabola, would not
seek, as our geometers do, to rectify the parabola, they would endea
vour, if one may coin the expression, to parabolify the straight line.'

The quotation above appears in Dr. J. F. Scott's ar

ticle "Boscovich1 s Mathematics," which is one of nine
articles in Roger Joseph Boscovich3 edited by Lancelot
Law Whyte and first published by George Allen & Unwin
in 1961.
ISSN 0-9771170-1-4