ARTICLE

The abstract algebra and the unsolved problems of ancient Greece

key words

algebraic structure of constructive numbers with ruler and compass, field , field extension, the degree
of field extension, the algebraic number, trisection of angle, doubling cube, squaring the circle

numbers and geometry in Greeks

Greek arithmetic and geometry were always together because they were considered two
different ways of exploring the same number system, so seemed very natural the geometric
constructions to perform arithmetic operations. Each number ought have a geometric construction.
And as the known numbers were the rational, the coexistence of arithmetic with the geometry was
harmonious. After the shock of the revelation that √2 was not rational, and the fact that he had a
geometric existence but non arithmetic 1, strengthened the belief in the geometric construction of
numbers. The numbers were constructible in geometry (of Euclidean genus) and vice versa only
numbers that would be presented in the geometrical construction could exist, meaning that we
construct geometrically the number a mean that construct the segment length |a|. Within the
framework of the three problems, our belief was that the numbers contained in them, clearly exist, so
should therefore be a corresponding geometric construction. But are all numbers constructible?
The geometry of the Greeks standardized by Euclid, was originally based on the concepts of point,
straight line and circle, and for this reason, tools for the study of geometry of trisecting the angle, the
doubling of the cube and squaring of circle was by plane methods alone, using straighetedge and
compass, for the construction of straight lines and circles. The three unsolved problems mentioned
above, raised in this Euclidean 'genus' of ' Elements' 2. All the constructions of this chapter will be
referred in this Euclidean genus.

Why does in this genus we could bisect an angle with ruler and compass, but no trisect it?
Or why we could construct a segment equivalent to α√2 ( α rational ) and not to a √
3
2? this
construction is difficult or impossible?
1 The √ 2 is the hypotenuse of a right triangle perpendicular sides 1.

2 The solution of Menechmos on the problem of doubling the cube using parabola and hyperbola, could not be drawn with a ruler and compass
 With the elementary theory of fields of abstract algebra, we can understand why the Greeks
were unable to solve the famous three problems, in Euclidean genus. The conceptual and
chronological distance between the geometry of Euclid and abstract algebra is huge, how they are
connected? With the common algebraic structure of constructible numbers, and that of field.

The answer will be given through the theory of the field extension.
In abstract algebra, we need some concepts for the three problems: the concept of field, of field
extension, of the degree of field extension, and of the algebraic number and irreducible
polynomials.

field extension
The set of rational Q is a field (satisfies some axioms, has a structure), if we add (attach) the
number √2 and all numbers with the operations of √2 with the rational, we have the  extension of Q,
→ Q(√ 2) containing numbers of the form a + b√ 2 with a, b rational .

The degree of this field extension is denoted [Q(√2):Q] is the degree of the irreducible
polynomial on Q, with √2 as a root i.e. x2 -2 (degree two, irreducible on Q).

Algebraic number, is any number that is the root of a polynomial with coefficients in Q, i.e.
the √2 is algebraic because it is the root of x 2 -2, o √
3
2 as a root of x3 -2 , e.t.c. Obviously the algebraic
numbers are extension of rational, any rational is algebraic. Every non- algebraic number is called
transcendental , i.e. the number e.

irreducible polynomial on Q: if the degree of a polynomial f is 2 or 3 then f is irreducible

in the field Q if and only if it has no roots in Q.

geometric image construction

the sets of points that are geometrically constructed are called constructible points. these are

1. assuming any two points O ( 0,0) and A (a, b), with coordinates rational numbers.
2 . construction of a straight line through two points or a circle from its center and radius

3.the intersection of two non parallels straight lines

4. the intersection of circles and straight line with cycles.

So produced all constructible points in a problem, since the shapes constructed by rule, and
compass are straight lines and circles.

Relations 2 and 3 reduce to rational operations. The intersection of a straight line and a
cycle or two cycles, are reduced to the solution of quadratic equation, i.e. the square root
extraction. So considering that the integers are easily constructible, it is known from high school
constructions to show that all the others.
geometrically constructible numbers
Let's take an example of geometrical construction of a number: the square root of
anypositiveintegerx.
As we have, we will construct a suitable orthogonal triangle
 where OA = (x-1/2 and OB = (x +1)/2 M midpoint of OB, then PB = B = x (the PB, PB’ are
tangents to the large circle so the OPB rectangle at P etc ...)
Here the constructible number √x, which corresponds to segment PB defined by a series of
constructible points A, B, P, P’ which emerged with the above
four possibilities provided by the rule, and compass.

The proof of equivalence between the algebraic and geometric

definitions of constructible numbers has the effect of
transforming geometric questions in Euclidean genus,
constructions into algebra.

constructible numbers, is a field in abstract algebra

this field is an extension of the rational numbers and it in turn is contained in the field of
algebraic numbers. It’s genesis is born from it’s properties:

A. rational numbers are constructible

B. if a > 0 is constructible then √a is constructible

C. if a,b are constructible similarly are and a ± b , a.b, a/b ( b ≠ 0) , i.e. the constructible
numbers are a field ( these are the definition), which contains the rational, is an extension of
rational since it contains √a.

theorem 1, if k is a constructible real number then k is the algebraic on Q , and the degree
of extension Q [( k )/ Q] is a power of 2

This theorem is a necessary condition for the existence of a constructible number k and
applying it, we can show that there is no geometrical construction for the three problems of antiquity
using only compass and ruler.

Trisection of angle. The first solution was from Hippias by using the squared curve. We show
that there is no always geometric construction for trisection of angle θ.
Knowing an angle is equivalent to know the cosine of the angle. So on to trisect the angle 3 θ we have
to construct the solution of (1): cos3θ = 4cos3θ – 3cosθ
If the angle θ=200 so cos3θ=1/2, the equation (1)becomes
8x - 6x -1 = 0 and the polynomial is irreducible on Q, with a real root a = cos20 and degree of
3 0

extension and [Q ( cos20 0 ):Q] =3, therefore cos20 0 is not constructible, i.e. the angle 20 0 is not
constructible(Theorem1.)
If the angle 3θ = 900 (1) becomes 4x3 - 3x = 0 which is not irreducible on Q , and also know that the 30
° angle is constructible.

Doubling the cube Solutions given by Hippocrates, Archytas, Menechmon etc. If x is the
edge of cube with twice the volume of the cube with edge 1 then
 . To construct, must be constructible the point (√ 2,0) But from the
3
x3 =2.13 → x=
theorem 1 appears impossible since√
3
2 is the root of the irreducible polynomial x3 -2 and thus [Q (√3 2
):Q]=3. (theor. 1).

Squaring the circle The problem is the construction of a square with an area equal to the
area of a circle 1. We investigate if the number √π is constructible. But from a classical result that
was shown by the F. Lindemann in 1882, we know that the number π(pi) is transcendental over Q,
i.e. it doesn’t satisfy some polynomial equation with rational coefficients ( not algebraic on Q,
Theorem 1 ). Hence √ π is transcendental too, so it is not constructible with ruler and compass.
Therefore, we cannot square the circle using only rule and compass.

Books :

“Introduction to Galois theory”:Adrew Baker (University of Glaskow)

“Algebraic extensions of fields” : Paul J. McCarthy Dover books

“Field and Galois theory” :J.S.Milne (inernet)

“Web sites” :Wikipedia, Wolfran Mathdord, Plan