Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
3 Descartes 6
3.1 Given the curve we get. . . . . . . . . . . . . . . . . . . . . . . . . 7
3.2 Problems with the method . . . . . . . . . . . . . . . . . . . . . . 9
4 Fermat 10
4.1 What the text says . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.1.1 Maxima and Minima . . . . . . . . . . . . . . . . . . . . . 10
4.1.2 Tangent . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
4.1.3 Center of Gravity . . . . . . . . . . . . . . . . . . . . . . . 13
4.2 The ideas behind . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.2.1 How to find the maximum . . . . . . . . . . . . . . . . . . 15
4.2.2 How to use the method for other problems . . . . . . . . . 16
4.3 Why it works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
4.3.1 The method for finding maxima . . . . . . . . . . . . . . . 17
4.3.2 The method for finding tangents . . . . . . . . . . . . . . . 17
4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2
1 Introduction
In the XVII century all of Europe is going trough a period of change, not only
in its way of living, but also, and even more deeply, in its way of thinking. A
big demographic, economic and political crisis happened in these years, whose
biggest evidences were the 30 years war (1618-1648), different civil and rural
revolts, the revolution and change of government in England, the expansion wars
carried out by Louis XIV (1638-1715), etc. The absolutism seemed to be so
much reinforced by this crisis that the XVII century has been presented as the
apogee of absolutism. The religion at this time is marked by the answer of the
Catholic Church to the Lutheran Reformation, with the Trento’s council. The
change is noted in all aspects of life, politics, science, religion, art. . . It’s the age
of rationalism and big thinkers. In the frame of politics is produced a transition
from the metaphysical and theological concept of the natural right, to a rational
concept. As well as the economics ones, the causes of this revolution are closely
connected to the progress of science and the discovery of new lands. The rights
kept separated of religion, and politics became independent of theology.
1.2 Mathematics
The development of other mathematical ideas at these times was highly connected
to the new conception of science. They turned to be the language of modern sci-
ence and its progress conditioned mathematics very much. There were born,
or renewed, different fields of mathematics: algebra, number theory, probability
3
calculus, projective geometry, infinitesimal calculus. . . The development of the
astronomy with its need of using and working with huge magnitudes and the
requirement of simplifying the calculus that the astronomic functions bring in-
volved the birth of the logarithms, invented by John Napier around 1614. There
were two principal ways of the mathematical activity in the XVII century. The
first one, based on “La Géometrie” (1637) of Descartes, was the introduction of
the analytic geometry. The second one was the invention of infinitesimal quan-
tities, with the publication in 1635 of “La Geometria” of Bonaventura Cavalieri.
We cannot forget the birth of number theory, and the application of algebraic
process to geometry and the creation of a new geometry, the projective geometry
by Gerard Desargues.
1.3 Mechanics
The science of mechanics was born in the XVII century. Its initiator was Galileo
(1564-1642). This new science essentially studies the laws of falling body and
the solution of the problem of the movement of a projectile without any resis-
tance of the environment. He experimented with a ball falling through a lightly
inclined plane and could establish the law of falling bodies, which says that the
distances are related with the square of the time. It is a new conception of
gravity in opposition with Aristotle’s conception. While Aristotle’s explanation
of the acceleration of the falling bodies is causal and qualitative, Galileo’s con-
ception comes from observation in different experiments, it is more quantitative.
He arranged to establish a law that defines the space in function of the time,
and he wanted to find out how to deduce this quantity logically from a simple
mathematical ratio. Galileo was creating the needed link between mathematics
and movement. In relation with the problem of the projectile, Galileo solved it
with the composition of motions, proving that the trajectory of projectile and
composition of horizontal and vertical movements is a parabola. If we take a look
further in the project, we’ll meet with Roberval and the cycloid, an example of a
curve obtained by composition of two motions. The method he uses for finding
the tangent is just adding the velocities of the two motions, isn’t it pretty related?
The ideas and the works of Galileo in mechanics were prolonged and well spread
by his disciples, among which is Torricelli. In mechanics also Descartes made
some important advances. His theory was very different from the other ones.
His mechanic lied on the construction of a complete system where a universal
mechanism explains all the phenomenon of the visible world. Descartes’ work is
the most important one in mechanics until the “Principia” of Newton. Although
the importance of Descartes’ work, the link between Galileo and Newton will be
set in by Christian Huygens’ work (1629-1695) . The most important of his work
was the discovery of the laws of centrifuge force.
4
1.4 Revolution of astronomy
The XVII century was a golden age for astronomy. An age when the conception of
Copernicus, which he had defended along the XVI century, will be confirmed and
completed. The characters that opened the door of the astronomy revolution
of the XVII century, were Kepler and Galileo. Johannes Kepler’s (1571-1630)
most important work was the establishment of the laws of gravitation of the
planets around the sun. As Kepler, Galileo was convinced of the exactitude of
the Copernican system. He studied the sky and his observations made clearer the
falseness of the sky immutability conception of Aristotle’s. He was able to give a
very good descriptions of the moon relief and some others discoveries, thanks to
a lens he had built in 1609. The culmination of the astronomy revolution of the
XVII century was represented by Newton works. He developed completely the
calculus of centripetal forces, deducted the movement of the planets, etc. He also
gave to astronomy a new instrument of observation: the reflexion telescope.
5
C1 C2 C3 C4 C5 A5
A4
A3
A2
A1
A R C
the vectors originates from the Greeks (Archimedes used kinematic concepts On
Spirals) and was exploited by Galileo and Torricelli.
3 Descartes
When Descartes in 1637 published his book La Gometrie he included a method
for finding a normal to a given curve. What Descartes wanted was to investigate
how different curves intersect with each others. He wanted to find the intersection
angle between curves by finding straight lines which are normal (perpendicular)
to the curves in the intersection point. Then he claimed that the intersection be-
tween two such straight lines will have the same angle as the intersection between
the two curves. Therefore he tried to make a method for finding such a normal
to a curve in a given point.
6
The idea of the method is to find a circle that touches the curve in a single
point and then, as the radius of the circle is the normal to the circumference of
it, we have the normal to the curve in the touching point.
Now, to find the desired point P , given the curve E, the point C and the
line of reference AG, assume that we have a point P on AG. Then P and C
determines a circle with P as its center and P C as radius. If the curve E cuts
the circle in anotherppoint than C, the equation for√the curve f (x, y) = 0, with x
substituted by v + s2 − y 2 or y substituted by ps2 − v 2 + 2vx − x2 will have
2 2
p for instance x by v + s − y in f (x, y) we get an
two roots. By substituting,
equation fe(y) = f (v + s2 − y 2 , y) = 0 in only one variable to solve for. If the
circle cuts the curve once, the equation must have one (double)root. Thus, what
we are looking for is to determine the coefficients s, v such that the equation
(with the substitutions) has a (double)root in C.
The way he shows the correspondence between the normal and the double
root is by assuming that a point P on AG gives a circle that cuts the curve in C
and another point C 0 . See figure 4. Then:
but if this point P be ever so little nearer or farther from A than
it should be, this circle must cut the curve not only at C but also
1
Descartes does not define the curve by a function, but he thinks of the points on the curve
as solutions to some equation(s). What we mean by f (x, y) = 0 is that a point Q is on the
curve iff the line AG has distance y to Q and the projection M of Q on AG has distance x to
A.
2
In [1] the x and y are swapped, but as it fits better into the modern notation to have x on
the horisontal axis and y on the vertical axis we have chosen to do so.
7
E
C
6
?
P M A
-
x -
v
in another point. Now if the circle cuts [the graph], the equation
involving x and y must have two unequal roots.
E
C0
0
A P P
0
G
Figure 4: If a point P is chosen different from P
The equation he refers to is the one determining the intersection between the
circle and the curve. Thus he does not really give a proof, but rather an intuitive
description of how it works. One can surely imagine curves not necessarily cutting
the circle twice. Later he claims that when moving the point P such that the
two roots gets closer and closer, they actually coincide, and in that situation the
curve touches, not cut the circle in the desired point:
8
coincide, the roots are exactly equal, that is to say, the circle through
C will touch the curve [E] at the point C without cutting it.
One may note that when writing this he is actually using the concept of
infinitesimal, but as we will see later it is not so. To find the point P such that
there is a double root in C one only uses the usually algebraic methods, but the
quote tells us that the idea to the method arrived from the notion of moving
points toward each others along a curve.
Lets say for instance, that the curve is given by the equation f (x, y) = x−y 2 =
2 3
0 and we wantp to find the normal to a point (x0 , x0 ). Then we must investigate
fe(y) = v + s2 − y 2 − y 2 = 0 or fb(x) = x − s2 + v 2 − 2vx + x2 = 0. If fb(x) is
to have a double root it must be on the form fb(x) = (x − x0 )2 = x20 − 2x0 x + x2
where x0 is the root. Therefore we have these three equations, as the coefficients
must be the same:
s2 − v 2 = x20
2v − 1 = 2x0
1=1
The second equation tells us that v = + x0 and thus the slope of the normal must
√ √ √
be − x0 /(v − x0 ) = 2 x0 and the slope of the tangent must be −1/(−2 x0 ) =
√
1/(2 x0 ). If fb(x) had been of degree greater than two we would be unable to
compare the coefficients directly, but we would know that fb(x) would have a
double root in x0 and thus be on the form fb(x) = (x − x0 )2 p(x) where p(x) would
be some polynomium with degree two less than fb(x). Hence we could compare
the coefficients, and hopefully solve the system of equations for v or s.
9
. . . This is my reason for believing that I shall have given here
a sufficient introduction to the study of curves when I have given a
general method of drawing a straight line making right angles with a
curve at an arbitrarily chosen point upon it. And I dare say that this
is not only the most useful and most general problem in geometry
that I know, but even that I have ever desired to know.
There is however a few problems in his method. If the curves gets just a little bit
complicated, the equations grow explosively and sometimes it is not even possible
to solve them.
For example the curve defined by f (x, y) = y 3 − x = 0 gives serious trouble.
by substitution one gets these two equations to work with:
√
s2 − v 2 + 2vx − x2 (s2 − v 2 + 2vx − x2 ) − x = 0
p
y 3 − v + s2 − y 2 = 0
4 Fermat
To solve any problem Pierre de Fermat (1601-1665) transforms it into the question
of finding the maximum or minimum of a polynomial. Then he uses his method of
adequating to find an equation that describes the maximum and thereby the solu-
tion to the problem. It is when adequating Fermat uses the concept of movements
and infinitesimals.
Source 1 consists of two letters. Both letters were sent to Father Marin
Mersenne. The first letter was sent in January 1638 and was intended for
Descartes. The second letter was sent in April 1638 and was intended for Rober-
val, but Father Mersenne also told Descartes of its contents.
We have divided the first letter in two parts so that the source is in three
parts: Maxima and Minima (from January 1638), Tangent (January 1638) and
Center of Gravity (April 1638).
After explaining what the text actually says, we will try to figure out what
the ideas behind the methods are and why the methods work.
10
Let a be any unknown of the problem (which is in one, two, or
three dimensions, depending on the formulation of the problem). Let
us indicate the maximum or minimum by a in terms which could be
of any degree.
A a E b−a C
Let E be the point that gives the maximum product. Fermat denotes AC
by b and AE by a so that the maximum is a(b − a). Then he looks at another
division of the line segment: a + e. The two products are then adequated and the
calculations made:
So the maximum appears in a = 2b . Fermat did not use “∼” but wrote “adequate”
(adégaler).
11
4.1.2 Tangent
Fermat then uses his method to find the tangent of a curve. He looks at a parabola
given by f (x, y) = 0 with f (x, y) = x − y 2 , and wants to find the tangent in a
point B. See figure 6.
B
O
C I D E
| {z }
e
| {z }
| d {z }
a
Let the tangent be given and let E denote the point where the tangent in-
tersects with the diameter of the parabola. Let D denote the point where the
parabola intersects with its diameter.
Fermat then looks at another point O on the tangent. Let C denote the
projection of B onto the diameter, and I the projection of O. The distance CI
is then the e from before.
Fermat has to make the problem of finding the subtangent CE into the ques-
tion of finding a maximum. He then establishes the following inequality:
CD CE 2
> .
DI EI 2
When establishing the inequality Fermat gives actual arguments:
CD = CB 2
DI < OI 2
12
because O is exterior to the parabola. The two triangles BCE and OIE are
similar, so
CB CE
=
OI EI
and thereby
CD CB 2 CE 2
> =
DI OI 2 EI 2
follows.
After having established the inequality, Fermat adequates the two sides
CD CE 2
∼
DI EI 2
and does the following calculations where a is the length of the subtangent CE
and d is the length of CD:
d a2
∼
d−e a2 + e2 − 2ae
da2 + de2 − 2dae ∼ da2 − a2 e
de2 + a2 e ∼ 2dae
de + a2 ∼ 2da
a2 = 2da
a = 2d
•
AI AN
=
AO AE
• M lies between N and I because of a property of the figure CBRV
13
A
b
E
O
N
B R
e M
C V
I
14
Now CAV − BAR = CBRV so we get that
ae
b OE
=
OM OM
CBRV
=
BAR
CAV BAR
= −
BAR BAR
IA2
= −1
AN 2
b2
= −1
(b − e)2
b2
= −1
b2 + e2 − 2be
2be − e2
=
b2 + e2 − 2be
which leads to the equation
b2 ae + ae3 − 2bae2
OM =
2b2 a − be2
From adequating OI ∼ OM and doing calculations similar to those of the
two previous examples, Fermat gets the result 2b = 3a.
15
x1 x2 a a+e
very close: they are adequated f (a) ∼ f (a + e), and then he divides through by e
before removing remaining appearences of e and putting the two equal. Fermat
sees no problem with this because the two values f (a) and f (a + e) are much
closer to each other than e is to zero: as Kepler states it
Near the maximum, the drecrements on both sides are initially im-
perceptible.6
CE 2 CD
−
EI 2 DI
6
Katz p. 470
16
which is of course entirely depended upon O. We see now that f (O) attains its
maximum in B. And that this maximum is 0.
Using the method to find the maximum, one would adequate
CE 2 CD
− ∼0
EI 2 DI
and then do some calculations ending up with an equation determining the max-
imum point B, i.e. an equation determining d by a. So in a way Fermat does
it the other way arround since it is a he is to determine. The method works,
however, because not two points on the upper half of the parabola have the same
subtangent.
Notice that Fermat does not look at f (d + e) but f (O) where O is B moved
on the tangent by e which is not the same as O + e.
Finding the center of gravity, Fermat uses the same approach. This time
finding the maximum of OM − OI.
f (a + e) f (a)
lim = lim .
e→0 e e→0 e
That is he sets f 0 (a) = 0 and then gets an expression that determines a.
Notice also that what Kepler said about
Near the maximum, the drecrements on both sides are initially im-
perceptible
df (a − e) (d − e)f (a)
lim = lim .
e→0 e e→0 e
17
Doing some calculations, we get df 0 (a) = f (a). √
Now placing the origo
√ so that
√ B has√coordinates ( √d, d), the √ tangent through
0
B is given by y = f ( d(x
√ √ − d) + f ( d), i.e. y = 2 d(x − d) + d. So a can
be determined as a = 2 d d = 2d.
Now, which functions satisfy that a = f 0 (f −1 (d))f −1 (d) exactly when df 0 (a) =
f (a)? These are the functions that Fermat’s method works on.
4.4 Summary
To solve any problem Fermat transforms it into the question of finding the maxi-
mum or minimum of a polynomial. Then he uses his method of adequating to find
an equation describing the maximum and thereby the solution to the problem.
Notice the idea of a point moving on the curve. When the method is used to
find tangents, it is a point on the tangent that moves towards the tangent point.
• The method for finding maxima is very efficient with a lot of polynomials.
The tangent method has a few demands to the polynomial f (x, y) = 0 of
the curve.
18
geometry. He didn’t even accepted curves which had not been described by some
movement first, while Fermat, once he had adopted algebra to geometry, defined
them in terms of equations. Descartes always described first the movement and
then he introduced the equation.
This view of moving points brought the idea of infinitesimal. Making points be
nearer and nearer is just making two movements be closer, what looks like a young
differentiation. And maybe you think that Descartes didn’t use infinitesimal, as
we have said he expressed his rejection to infinitesimal, and he didn’t work with
them as an explicit tool, as Fermat did, but the truth is that he developed his
method from a reasoning based in this infinitesimal, in this idea of moving points
along a curve, making them be very close to each other. So the question is:
infinitesimals as an explicit tool or just as a way of thinking? and we take the
risk and we go a little further, Is it really different?
19
of finding tangents of the polynomial f (x, y) = 0, which can be found in a letter
to Henry Oldenburg. They provided thus general algorithms for curves given by
polynomial equations. However this improvement in the computations wasn’t
acompanied with any change in the way of thinking, neither of infinitesimal nor
of the inverse relation between tangents and areas.
I R
P D
V
G E
Figure 9:
See figure 9. First take any curve ZGE whose perpendicular ordinates con-
tinually increase (V Z < P G < DE) and consider its “integral” V IF , that is: if
R is given, then the area V DEZ should be equal to DF · R.
20
DE R
See figure 10. Now the claim is that if DF
= DT
then T F is the tangent at F
D
V
T
Z
Figure 10:
of V IF .
See figure 11. To prove it, take any point I on the curve V F (first between V
I
K
T P D
Figure 11:
and F ). We get:
LF DF
= (similar triangles)
LK DT
DF DE
= (by the claim)
DT R
See figure 12. Which gives
RLF = LKDE
21
F
K L
T D
Figure 12:
P D
V
G
E
Figure 13:
22
I
L
K
T D
Figure 14:
y
F (x)
t(x)
a x0 x
−(1 + f (x))
Figure 15:
F (x0 ) 1
F 0 (x0 ) = = f (x0 )
x0 − a R
which is nothing else than the first statement of the fundamental theorem of
calculus Z x
( f (y)dy)0 = f (x)
0
23
6.3 Back to the historical connection with Leibniz and
Newton
In the proof he doesn’t need anymore to describe a curve by motions or in an
algebraic way. He just states pointwise the existence of such a height, which, as for
the length DT , relies on the completeness of R. As important and fundamental
his results might be, Isaac Barrow wasn’t able to translate these results into a
method of computation which could have been used for until then unsolvable
problems.
• they introduced their own notations and algorithms allowing an easy use of
the different concepts
• they understood and applied the inverse relationship between tangents and
areas
References
[1] Descartes; source 3
[3] Carl B. Boyer, The history of the calculus and its conceptual development,
second edition, Dover publications 1959.
8
We don’t know exactly if Newton actually followed any of Barrow’s lectures
24
[4] David M. Burton, The history of mathematics, an introduction, third edition,
McGraw Hill, 1997.
[9] www.artehistoria.com
25