HISTORY OF MATHEMATICS

Medieval Period (Continuation)

The
Development
of Calculus
HISTORICAL BACKGROUND by BURTON

Ancient Greek mathematicians — most notably Eudoxus and Archimedes — had successfully studied some
of the basic problems and ideas from integral calculus, and their method of exhaustion was similar to the
modern approach in some respects (successive geometric approximations using figures whose
measurements were already known) and different in others (there was no limit concept, and instead there
were delicate reductio ad absurdum arguments).

Towards the end of the 16th century there was further interest in the sorts of problems that Archimedes
studied, and later in the 17th century this led to the independent development of calculus in Europe by Isaac
Newton (1643 – 1727) and Gottfried Wilhelm von Leibniz (1646 – 1716).

Two important factors led to the development of calculus originated in the late Middle Ages, and both
represented attempts to move beyond the bounds of ancient Greek mathematics, philosophy and physics.

1. Interest in questions about infinite processes and objects. Ancient Greek mathematics
was not equipped to deal effectively with Zeno’s paradoxes about infinite and by the time of
Aristotle it had essentially insulated itself from its “horror of the infinite.” However, Indian
mathematicians did not share this reluctance to work with the concept of infinity, and Chinese
mathematicians also used methods like infinite series when these were convenient.

During the 14th century the mathematical work of Oresme and others on infinite series
complemented the interests of the scholastic philosophers; questions about the infinite played
a key role in the efforts of scholastic philosophy to integrate Greek philosophy with Christianity.
Philosophical writings by scholars including William of Ockham (1285 – 1349), Gregory of
Rimini (1300 – 1358), and Nicholas of Cusa (1401 – 1464) provided insights which anticipated
the concept of limit — a basic part of calculus as we know it.

2. Interest in questions about physical motion. During the 13th century Jordanus de Nemore
discovered the mathematically correct description for the physics of an inclined plane (a result
that had eluded Archimedes and was described incorrectly by Pappus), and later members of
the Merton school in Oxford such as T. Bradwardine (1290 – 1349) and W. Heytesbury (1313
– 1373) had discovered an important property of uniformly accelerated motion; namely, the
average velocity (total distance divided by total time) is in fact the mathematical average of
the initial and final velocities. The significance of this concept for the motion of freely falling
bodies was not understood at the time and would not be known until the work of Galileo and
others in the 16th century.
 Oresme’s writings on graphs contain an early insight that the area under a graph of velocity
with respect to time is equal to the total distance traveled.

During the late 16th century interaction between mathematics and physics began to increase
at a much faster rate, and many important contributors to mathematics such as S. Stevin also
made important contributions to physics (in Stevin’s case, the centers of mass for certain
objects and the statics and dynamics of fluids).

Problems leading to the development of calculus

In the early 17th century mathematicians became interested in several types of problems, partly because
these were motivated by advances in physics and partly because they were viewed as interesting in their
own right.

1. Measurement questions such as lengths of curves, areas of planar regions and surfaces,
and volumes of solid regions, and also finding the centers of mass for such objects.

2. Geometric and physical attributes of curves, such as tangents and normals to curves, the
concept of curvature, and the relation of these to questions about velocity and
acceleration.

3. Maximum and minimum principles; e.g., the maximum height achieved by a projectile in
motion or the greatest area enclosed by a rectangle with a given perimeter.

Specific examples of all three types of problems had already been studied by Greek mathematicians. The
work of Zenodorus (200 – 140 B.C.E.) on problems in the third type because it is reproduced in Pappus’
anthology of mathematical works (the previously cited Collection or Synagoge). Some examples of
Zenodorus’ results are

(1) among all regular n – gons with a fixed perimeter, a regular n – gon encloses the greatest
area,
(2) given a polygon and a circle such that the perimeter of the polygon equals the circumference
of the circle, the latter encloses the greater area,
(3) a sphere encloses the greatest volume of all surfaces with a fixed surface area.

Usually, the formula for the volume of a cone is derived only for a right circular cone in which the axis line
through the vertex or nappe is perpendicular to the plane of the base. It is naturally to ask whether there is
a similar formula for the volume of an oblique cone for which the axis line is not perpendicular to the plane of
the base (see the drawing below). One can answer this question using an approach developed by B.
Cavalieri (1598 – 1647). Similar ideas had been considered by Zu Chongzhi, also known as Tsu Ch’ung –
chin (429 – 501), but the latter’s work was not known to Europeans at the time.

CAVALIERI’S PRINCIPLE. Suppose that we have a pair three-dimensional solids S and T that lie
between two parallel planes P1 and P2, and suppose further that for each plane Q that is parallel to the
latter and between them the plane sections Q ∩ S and Q ∩ T have equal areas. Then the volumes of S
and T are equal.

Here is a physical demonstration which suggests this result: Take two identical decks of cards that are neatly
stacked, just as they come right out of the package. Leave one untouched, and for the second deck push
along one of the vertical edges so that the deck forms a rectangular parallelepiped as below.
In this new configuration the second deck has the same volume as first and it is built out of very thin
rectangular pieces (the individual cards) whose areas are the same as those of the corresponding cards in
the first deck. So, the areas of the plane sections given by the separate cards are the same and the volumes
of the solids formed from the decks are also equal.

We shall now apply this principle to cones. Suppose we have an oblique cone as on the right-hand side of
the figure below. On the left-hand side suppose we have a right circular cone with the same height and a
circular base whose area is equal to that of the elliptical base for the second cone.

Figure illustrating Cavalieri’s Principle for cones

In the notation of Cavalieri’s Principle, we can take P1 to be the plane containing the bases of the two cones
and P2 to be the plane which contains their vertices and is parallel to P1. Let Q be a plane that is parallel to
both of them such that the constant distance between P2 and Q is equal to y; we shall let h denote the
distance between P1 and P2, so that h is also the altitude of both cones. If b denotes the areas of the bases
of both cones, then the areas of the sections formed by intersecting these cones with Q are both equal to b
y2/ h2. Therefore, Cavalieri’s Principle implies that both cones have the same volume, and since the volume
of the right circular cone is b h2/ 3 it follows that the same is true for the oblique cone.

One approach to phrasing our physical motivation in mathematical terms is to imagine each cone as being a
union of a family of solid regions given by the plane sections; for the right circular cone these are regions
bounded by circles, and for the oblique cone they are bounded by ellipses. Suppose we think of these
sections as representing cylinders that are extremely thin. Then in each case one can imagine that the
volume is formed by adding together the volumes of these cylinders, whose areas — and presumably
thicknesses — are all the same, and of course this implies that the volumes of the original figures are the
same. One fundamental question in this approach is to be more specific about the meaning of “extremely
thin.” Since a planar figure has no finite thickness, one might imagine that the thickness is something less,
and this is how one is led to the concept of a thickness that is infinitesimally small.

Mathematicians and users of mathematics have thought about infinitesimals for a long time. They already
appear in the mathematics of the early Greek atomist philosopher Democritus (460 – 370 B.C.E.) and an
approach to squaring the circle developed by Antiphon, but advances by Eudoxus and others during the 4th
century B.C.E. enabled Greek mathematicians to avoid the concept, and this fit perfectly with the reluctance
of Greek mathematicians and philosophers (e.g., Aristotle) to eliminate questions about the infinite from their
mathematics.

Taking the somewhat obscure form of “indivisibles,” they reappeared in the mathematics of the late Middle
Ages, and they played an important role in the work of J. Kepler (1571 – 1630) on laws of planetary motion,
particularly his Second Law which states that the orbits of planets around the sun sweep out equal areas
over equal times.
During the 17th century infinitesimals were used freely by many mathematicians and scientists who
contributed to the development of calculus, and in particular both Newton and Leibniz used the concept in
their definitive accounts of the subject. However, as calculus continued to develop, doubts about the logical
soundness of infinitesimals also began to mount. Such questions ultimately had very important
consequences for the development of mathematics.

Progress on measurement questions

Methods and results from Archimedes and others provided important background and motivation for work in
this area. In particular, during the time before the appearance of Newton’s and Leibniz’ work, many of the
standard examples in integral calculus had been worked out by preliminary versions of methods that
became standard parts of the subject. Here are some specific examples:

1. Integrals of polynomials and more general power functions.

Cavalieri computed the integral of the function x n geometrically in cases where n is a positive
integer. Subsequently G. de St. – Vincent (1584 – 1667) integrated 1/x in geometric terms that are
equivalent to the usual formula involving log e x, and J. Wallis (1616 – 1703) generalized the
integral formula for x n to other real values of n.

Wallis was a particularly important figure in the development of calculus for several reasons. His
methods, replaced geometric techniques with algebraic computations and analytic considerations,
and as such they are a milestone in the development of analysis (calculus) as a subject distinct from
both algebra and geometry. In an entirely different direction, Wallis is also known today for his
applications of integral formulas to derive his infinite product formula for π/2:

We should note here that an earlier infinite product formula involving π

had been discovered by Viète (and, as noted earlier, Indian mathematicians had already discovered
some of the standard infinite series formulas involving π).

2. Measurements involving the cycloid curve and regions partially bounded by it. There was an
enormous amount of interest in the properties of the cycloid curve during the 17th century, and there were
also some bitter disputes about priorities among some of the numerous mathematicians who worked on
this example. Results included computations of arc lengths as well as areas, volumes and centers of
mass associated to the curve. Mathematicians associated with this work include B. Pascal (1623 – 1662,
better known for numerous other contributions to mathematics, physics and philosophy), E. Torricelli
(1608 – 1647, best known for work on the physics of fluids), and G. P. de Roberval (1602 – 1675).

3. Infinite solids with finite volume. Torricelli discovered a fact that few if any others had anticipated;
namely, the existence of an unbounded solid of revolution whose volume is finite. His example is an
unbounded piece of the solid formed by rotating the standard hyperbola y = 1/ x about the x – axis
(sometimes known as Torricelli’s trumpet).
 (Source: http://www.coopertoons.com/caricatures/torricellistrumpet.html )

4. Arc length. Early in the 17th century there were doubts about the possibility of computing the arc lengths
of many curves, including some extremely familiar examples. Results of H. van Heuraet (1633 – 1660)
showed the problem of finding arc length of a given curve is equivalent to determining the area under
another curve, and he also worked out certain examples including the semi – cubical parabola a y 2 = x 3.
An example of such a curve is given below (in red).

(Source: http://mathworld.wolfram.com/SemicubicalParabola.html)
The arc length of a spiral curve was computed by Roberval.

5. Integrals and series expansions of transcendental functions. Results on the integrals of the
standard trigonometric functions were obtained by Pascal, Roberval, Isaac Barrow (1630 – 1677) and J.
Gregory (1638 – 1675). The standard infinite series expansion for arctan x was obtained by Gregory
(however, the Indian mathematician Madhava had discovered the formula three centuries earlier), and
the standard infinite series expansion for log e (1 + x) was obtained independently by W. Brouncker (1620
– 1684) and N. Mercator (1620 – 1687). Gregory also made numerous other contributions, including
extending and applying the classical method of exhaustion to questions about other conic sections and
writing the first text covering the material that would become calculus.

Progress in differentiation and maximization/minimization problems

Greek mathematics provided far less insight into questions about tangent lines or maximizing functions than
it did for computing lengths, areas and volumes. In particular, there were no general principles comparable
to the method of exhaustion for describing tangents; each example was treated in an entirely separate
manner. Therefore, one major problem facing 17th century mathematicians was to produce a workable
concept of tangent line.

Some names particularly associated with this problem are Descartes, Fermat, Roberval and Barrow.
Descartes’ approach was based on finding the normal (perpendicular) line to a curve at a point, and
Roberval’s was motivated by the standard interpretation of a parametrized curve as the path of a moving
object. Fermat and Barrow both defined tangent lines by the method that became standard; namely, the
tangent line to a curve C at a point x is a limit of the secant lines joining x to a second point y on
C as y approaches x (see http://www.youtube.com/watch?v=gvCG_8CSBfo for an animated picture). Fermat
understood the basic idea, but Barrow’s language (i.e., his differential triangle) was more precise.
As in the case of measurement problems, examples were the focal point of work on tangents. The standard
results on the slopes of tangent lines for polynomial graphs were obtained by Fermat in the monomial case
and in general by Hudde. Applications of derivatives to repeated polynomial roots were also discovered at
this time. Tangent lines to the cycloid were determined independently by Descartes, Fermat and Roberval.

Fermat also studied maximization and minimization problems using the approach he developed for tangent
lines.

One result of Kepler’s on maximization (the Wine Barrel Problem). He observed that a wine merchant had
figured the amount of wine in a barrel by inserting a rod into the barrel diagonally through a small hole in the
top. When the rod was removed, the length L of the rod which was wet determined the price for the barrel.
Kepler was concerned about the uniformity of pricing by this method and decided to analyze its accuracy; he
correctly realized that a taller, narrower barrel might yield the same measurement as a shorter, wider one,
resulting in the same wine price, even though its volume would be considerably smaller (see the figure below).

(Source: http://www.ugrad.math.ubc.ca/coursedoc/math100/notes/apps/maxmin.html )

In his approach to determine the volume, Kepler approximated the barrel by a cylinder with base radius r of
the base and height h; the conditions of the problem imply the equation r 2 + h2 = L 2. He then looked for
the values of r and h giving the largest volume V when L (which determines the cost) is held constant.
Using differential calculus one can show that the relation between L and h must be L 2 = 3 h2, which is
the answer that Kepler found using less refined methods. He also observed that the shapes of the wine
barrels were close to this optimal value — in fact, so close that he could not believe this was a coincidence.
Of course, the manufacturing processes then were less uniform than they are now, so is was unlikely that all
barrels satisfied this mathematical specification precisely, but Kepler further noted that if a barrel deviated
slightly from the optimal ratio this would have little effect on the volume because the volume function changes
very slowly near its maximum.

Other contributions of Kepler

He is also known for stating his so – called Kepler sphere packing problem, which was first formulated in
1611 and conjectures that the most efficient way of packing solid spheres into a box is the usual
method in which oranges (or other solid round objects) are stacked on top of each other (see the figures
below).
 (Source: http://www.math.binghamton.edu/zaslav/Nytimes/+Science/+Math/sphere-packing.20040406.gif)

The emergence of calculus

Since many of the basic facts in calculus were known before the work of Newton and Leibniz, it is natural to
ask why they are given credit for inventing the subject. Others came very close to doing so; in particular,
Barrow understood that the process for finding tangents (differentiation in modern terminology) was inverse
to the process for finding areas (integration in modern terminology). In these notes we shall focus on the
decisive advances that make the work of Newton and Leibniz stand out from the important, high-quality
results due to many of their excellent and talented contemporaries.

1. Before Newton and Leibniz techniques for differentiating and integrating specific examples
had been developed, but they were the first to set general notation and define general
“algorithmic processes” for each construction. Earlier workers were not able to derive useful
and general problem – solving methods.

2. Newton and Leibniz recognized the usefulness of differentiation and integration as general
processes, not just as ad hoc methods to solve measurement and tangent problems in
important special cases. Apparently, no one (at least in Europe) had previously recognized
the usefulness of calculus as a general mathematical tool.

3. With the exception of Barrow, the mutually inverse relationship between differentiation and
integration had not been clearly recognized in earlier work, and Newton and Leibniz were the
first to formulate it explicitly and establish it in a logically convincing manner.

4. Both stated the main ideas and results of calculus algebraically, so that the subject was no
longer an offshoot of classical Greek geometry but significantly broader in scope and poised
to make fundamental contributions to many areas of knowledge.

5. Both were able to go much further than the Kerala school in India and K. Seki in 17 th century
Japan; neither of the latter had the advantage of newly formulated problems from recent and
extensive scientific investigations, and the questions arising from these studies were crucial
to determining the breadth and generality in the work of Newton and Leibniz.
The discoveries of Newton and Leibniz were essentially independent, and although Newton was the first to
develop the subject, Leibniz published his version first.

The discoveries by Newton and Leibniz took place around 1665 and 1673 respectively. Leibniz’ work was
published in 1684 while Newton’s was published 1736, nearly a decade after his death.

There are many similarities why Newton and Leibniz are given credit for creating calculus. One similarity is
that each used both differentiation and integration to solve difficult and previously unanswered questions.
Both also proved many of the same basic results; e.g., the standard rules for differentiating functions, the
Fundamental Theorem of Calculus, and the basic formal integration techniques which appear in calculus
textbooks. On the other hand, Newton and Leibniz clearly had different priorities and these can be seen in
some differences between their approaches and conclusions.

1. The standard binomial series expansion for (1 + x) a, where a is an arbitrary real number and
| x | < 1, is solely due to Newton.

2. Newton used the words fluxion and fluent to denote the derivative and integral, and he denoted
derivatives by placing dots over variables.
Leibniz ultimately adopted the dx notation and the integral sign that have been standard for
centuries.

3. Newton was primarily interested in the uses of calculus to study problems involving motion,
while Leibniz’ work and interests involved finding extrema and solving differential equations.

4. Newton discovered the rules and processes of calculus by a study of velocity and distance,
while Leibniz did so via algebraic sums and differences.

5. Newton used infinitesimals as a computational means, while Leibniz used them directly.

6. Newton’s priority was differentiation while Leibniz’ was integration.

7. Newton stressed the use of infinite series to express functions, while Leibniz preferred
solutions that could be written in finite terms.

8. Leibniz gave more general rules and more convenient notation.

Despite the clear differences in the Newton and Leibniz approaches to calculus, each perspective has
advantages in certain situations. For example, some functions cannot be expressed in finite terms
involving standard functions from first year calculus but can be studied very effectively using infinite series
as in Newton’s treatment, while others cannot be studied using infinite series and thus are more compatible
with Leibniz’ viewpoint.

The other important writings of Newton and Leibniz reflect some of the differences mentioned above. Leibniz
wrote many lengthy and highly influential works on philosophy (which as usual are beyond the scope of this
course), while Newton wrote several important books on the sciences. The latter include his best-known
work, Philosophiæ naturalis principia mathematica (The Mathematical Principles of Natural Philosophy),
which remains of the most important books in the sciences ever written. In this book he developed the laws
of motion using calculus and used them to derive Kepler’s laws of planetary motion. E. Halley (1656 – 1742,
the namesake of the famous comet) played a major role in persuading Newton to publish a complete version
of his work
Differences of Newton’s and Leibniz’ legacies with respect to calculus

1. Newton’s applications of calculus ultimately determined the direction of subsequent work in

mathematics and physics.

2. Leibniz’ formulation of calculus ultimately determined how the mathematical aspects of this work
were formulated (however, Newton’s dot notation for derivatives is still used sometimes in
physics to denote derivatives with respect to time).

In other words, Newton’s legacy is more about the sorts of scientific problems that
calculus has considered during the past three to four centuries, while Leibniz’ legacy
is more about the way such problems are studied.

Other contributions of Newton and Leibniz

Both Newton and Leibniz made other substantial contributions to mathematics aside from their momentous
work on calculus.

Newton — Co-discovered (with J. Raphson, 1648 – 1715) the Newton – Raphson method for
numerically approximating solutions to complicated nonlinear equations, including irrational roots of
higher degree polynomials. This is one of the best-known methods for finding such solutions; there
are descriptions of this method in many (most?) standard calculus textbooks.

Newton — In an appendix to his major scientific work titled Opticks (published in 1704), he classified
all curves in the coordinate plane which are defined by third degree polynomial equations in the
coordinates x and y. This has been characterized as the first comprehensive study of curves and
their properties since the work of Apollonius. Newton also did several other important pieces of
algebraic work, including his generalization of the binomial formula and his co-discovery (with A.
Girard) of certain important identities involving symmetric polynomials in n algebraically independent
variables (a polynomial p(x 1, ... , x n) is said to be symmetric if any permutation of the variables
yields the same polynomial; for example, the sum x 1 + ... + x n and product x 1 ... x n are symmetric
but the polynomials x 1 and x 1 x 3 + x 2 x 3 are not).

Newton — Contributions to the calculus of finite differences, whose uses include describing
recursively defined sequences like the Fibonacci numbers.

Leibniz — Early work on studying logic using algebraic methods to manipulate well – formed
grammatical statements. This has been described as one of the most important advances in the
theory of logic between the time of Aristotle and the more systematic work on symbolic logic in the
19th century by G. Boole (1815 – 1864) and others.

Leibniz — A comprehensive formulation of the binary or base 2 numeration system. This expanded
upon many earlier ideas about representing numbers or words by combinations of two symbols,
including numeration systems in a few cultures and the ancient writings of Pingala from 2000 years
earlier, and in fact Leibniz explicitly cited symbolism in the ancient Chinese I Ching (in modern Pinyin
transliteration, Yì Jīng) as one forerunner of his work.

Leibniz — Solutions of systems of linear equations, including the definition of the determinant of a
square matrix and the standard formula for it as a polynomial function of its entries (the Japanese
mathematician K. Seki did similar things independently, and in virtually complete isolation, about one
decade earlier).

Compiled by: Russel Paud