Discover millions of ebooks, audiobooks, and so much more with a free trial

Only $11.99/month after trial. Cancel anytime.

Geometry of Complex Numbers
Geometry of Complex Numbers
Geometry of Complex Numbers
Ebook361 pages3 hours

Geometry of Complex Numbers

Rating: 4 out of 5 stars

4/5

()

Read preview

About this ebook

"This book should be in every library, and every expert in classical function theory should be familiar with this material. The author has performed a distinct service by making this material so conveniently accessible in a single book." — Mathematical Review
Since its initial publication in 1962, Professor Schwerdtfeger's illuminating book has been widely praised for generating a deeper understanding of the geometrical theory of analytic functions as well as of the connections between different branches of geometry. Its focus lies in the intersection of geometry, analysis, and algebra, with the exposition generally taking place on a moderately advanced level. Much emphasis, however, has been given to the careful exposition of details and to the development of an adequate algebraic technique.
In three broad chapters, the author clearly and elegantly approaches his subject. The first chapter, Analytic Geometry of Circles, treats such topics as representation of circles by Hermitian matrices, inversion, stereographic projection, and the cross ratio. The second chapter considers in depth the Moebius transformation: its elementary properties, real one-dimensional projectivities, similarity and classification of various kinds, anti-homographies, iteration, and geometrical characterization. The final chapter, Two-Dimensional Non-Euclidean Geometries, discusses subgroups of Moebius transformations, the geometry of a transformation group, hyperbolic geometry, and spherical and elliptic geometry. For this Dover edition, Professor Schwerdtfeger has added four new appendices and a supplementary bibliography.
Advanced undergraduates who possess a working knowledge of the algebra of complex numbers and of the elements of analytical geometry and linear algebra will greatly profit from reading this book. It will also prove a stimulating and thought-provoking book to mathematics professors and teachers.

LanguageEnglish
Release dateMay 23, 2012
ISBN9780486135861
Geometry of Complex Numbers

Related to Geometry of Complex Numbers

Titles in the series (100)

View More

Related ebooks

Mathematics For You

View More

Related articles

Reviews for Geometry of Complex Numbers

Rating: 3.8333333333333335 out of 5 stars
4/5

6 ratings0 reviews

What did you think?

Tap to rate

Review must be at least 10 words

    Book preview

    Geometry of Complex Numbers - Hans Schwerdtfeger

    NUMBERS

    INTRODUCTION

    NOTE ON TERMINOLOGY AND NOTATIONS

    IN MOST branches of mathematics, including the subject to be dealt with here, a general agreement concerning terminology and notation has not been reached and indeed may not be desirable. The author has tried, however, to introduce such names and terms as have been adopted in the majority of writings on the geometry of complex numbers; he has avoided terms suggesting a special interpretation which he did not wish to emphasize. Thus we speak consistently of ‘circles,’ avoiding the term ‘chain.’

    The present exposition starts on the basis of analytical geometry using cartesian coordinates x, y of the points in the plane. These coordinates are any real numbers. For our purposes it is expedient to represent the point (x, y) by the complex number z = x + iy where i = √(–1). We call x = Re z the real part, y = Im z the imaginary part of z = x iy is the conjugate, and |z| = √(x²+y²) = r 0 is the modulus or absolute value of z. We notice that z = r² is zero if and only if z = 0. The elementary algebra of the complex numbers is assumed to be known.¹

    Using polar coordinates r, θ we may represent the complex number z in the ‘trigonometrical form’

    z = r(cos θ+isin θ).

    We shall call cos θ + i sin θ the ‘angular factor’ of z. It is a complex number of modulus one, commonly denoted by eiθ. This exponential notation is justified by the identity

    e1 e2 = ei(θ1+θ2),

    that is, the functional equation of the exponential function, which is a formal consequence of the addition theorems of elementary trigonometry.² Knowledge of the theory of functions of a complex variable will not be assumed ; however, we shall not explain the concepts of continuity and differentiation, which will be used occasionally.

    Instead of the angular factor eiθ of a complex number z we shall sometimes have to introduce the angle θ itself as a function of z ; it will be denoted by

    θ = arc z = arc ei(θ +²kπ )          (k integral).

    It is defined for all z ≠ 0, up to an additive integral multiple of 2π. This multiple may be chosen conveniently from the provenance of the number z. But the range of the values of arc z is often restricted to a certain interval of length 2π, for example, –π < θ π (the principal value of arc zθ < 2π; the function arc z is then discontinuous along the negative or the positive real axis (x-axis) respectively. Its jump value along this axis is ± 2π depending upon the direction in which z crosses the axis.

    The basic facts of Linear Algebra are required to an extent not exceeding the amount necessary for an appreciative understanding of the elements of analytic geometry,³ including homogeneous coordinates (L.A. Appendix, no. 1). We shall make extensive use of the formal laws of matrix algebra (L.A. §§ 12–13); but in the initial stages we shall use only two-rowed matrices whose elements are complex numbers. Such matrices will be denoted by German capitals, for instance

    stands at the same time for the matrix, as well as for the Moebius transformation

    is defined apart from a complex number factor q and q representing one and the same Moebius transformation, could have been identified; we have preferred, however, to use only strict matrix equations although in many cases the factor q may remain undetermined.

    (λ .’

    The determinant ׀.

    Transformations play an important role in our exposition. A transformation Z = f(z) is given by a complex-valued function f(z) which associates with every point z in the plane of the complex numbers (or in a certain part of this plane) a complex number Z that appears as a point in the same plane, or in another plane with a congruent coordinate system. We also say that the transformation carries the point z into the point Z (cf. L.A. § 10).

    In many cases a set of transformations relevant for a geometrical discussion forms a group (cf. L.A. § 22) where the group multiplication is the composition of two transformations of the given set in the sense of forming the function of a function. The unit element in such a group is the identity transformation Z = z which relates every point with itself. Apart from a certain familiarity with the general idea of a group which may be gained by dealing with the special groups of transformations occurring in this book, no elaborate knowledge of group theory is required.

    The systems of all rational, all real, or all complex numbers are fields in the sense of algebra. This means that within each of these systems the operations of elementary arithmetic—addition, subtraction, multiplication, and division (except by zero)—can be performed so that the sum, difference, product, and quotient of any two numbers of one of these systems are again numbers of the same system. There are many other number fields beside those mentioned above; all of them are contained in the field of all complex numbers and contain in their turn the field of the rational numbers as a subfield (cf. L.A. § 14). In one instance only the reader should realize that fields exist, the elements of which are not numbers, for example, fields consisting of a finite number of elements (L.A. ex. 16.1). In such a field the elementary operations are defined so that the formal laws of arithmetic, namely the existence and uniqueness of the zero and unit elements, as well as negatives and inverses, and the associative, commutative, and distributive laws, are satisfied.

    2 are said to be isomorphic (L.A. § 14) if their elements a1, b1,…, and a2, b2,…, can be put in one-one correspondence, a1 ↔ a2, b1 ↔ b2,…, so that a1 + b1 ↔ a2 + b2, a1b1 ↔ a2b2 respectively.

    2 represent the same field) is called an automorphism. All automorphisms of a field form a group. For the field of all rational and the field of all real numbers this group consists of the identity only: a ↔ a (L.A. ex. 14.8). In the field of all complex numbers the correspondence z ↔ ; is readily seen to be an automorphism different from the identity.

    Primitive concepts and facts of topology will be used. For explanations and proofs we refer to the literature, mainly books containing introductory chapters.⁴ For some examples in chapter III, line and area integrals are assumed to be known. References will be given.

    Finally we may point out a number of embarrassing features in the terminology; the difficulties are mainly linguistic in so far as different matters are described conventionally by the same word.

    There are different meanings attached to each of the words ‘real’ and ‘imaginary’ in geometry. The real numbers are the complex numbers associated with the points of (or lying on) the real axis of the complex plane, also called the Gauss plane (or the Argand plane). The pure imaginary numbers are in the same way associated with the real points of the imaginary axis. (Point and associated complex number will always be denoted by the same symbol.) All points in the (complex number, or z-) plane, each characterized by a complex number, are ‘real points.’ Apart from figures consisting of real points we shall have to consider (only as purely algebraic creations) certain figures, for example, ‘imaginary circles’ (also called ‘ideal circles,’ cf. Deaux [1], p. 49) whose (non-real) ‘points’ have non-real cartesian coordinates; these imaginary circles admit of a real geometrical interpretation only by means of the associated inversion (cf. § 2). Single ‘imaginary points’ will not occur. Also observe that real circles have imaginary points. A real circle may have a non-real (imaginary) intersection with the real axis; thus there are even non-real points on the real axis!

    A word may be said about the use of the adjectives ‘hyperbolic,’ ‘parabolic,’ and ‘elliptic.’ As a matter of fact, there is little justification for these except the reference to the various kinds of non-euclidean geometry where their use is by now historic. So we have hyperbolic, parabolic, and elliptic pencils and bundles of circles as well as motions. But there are also hyperbolic, parabolic, and elliptic Moebius transformations, inversions, and antihomographies, to mention only such objects as will occur in the present book. In these cases the relation to the classical terminology in non-euclidean geometry is more or less remote. General agreement on these definitions seems to exist, however; thus we have accepted them apart from one minor alteration to be explained in § 8, c, footnote 4.

    ¹ Cf. R. Courant and H. Robbins [1], chapter II, § 5; or G. H. Hardy [1], chapter III.

    ² Cf. Courant–Robbins [1], pp. 477–9; or Hardy [1], §§ 232–3.

    ³ Cf. the early chapters of any text on Linear Algebra, in particular the author’s [2], in the following quoted as ‘L.A.’

    ⁴ Caratheodory [3], vol. I, Part II. Courant and Robbins [1], chapter v. Hilbert and Cohn-Vossen [1], chapter VI.

    I

    ANALYTIC GEOMETRY OF CIRCLES

    §

    1. Representation of circles by hermitian matrices

    a. One circle. All points z = x+iy of the complex plane which form the circumference of the circle of radius ρ about the centre γ = α + iβ are characterized by the equation

    (x–α)² + (y–β)² = ρ².

    we obtain the equation of the circle in the form

    For our purposes it is advisable to start with the more general equation

    where A and D are real, and B and C are conjugate complex numbers. The matrix

    is therefore a hermitian matrix.

    Evidently the equation (1.1) is a particular case of (1.2). Both equations will represent the same circle if A ≠ 0 and

    (1.3)  

    is associated with an equation (1.2). We shall say that it defines, or is representative of, a ‘circle’ except if A = B = C where λ is a real number different from zero.

    In order to distinguish between the different types of ‘circles’ included in this definition we introduce the determinant

    | = AD–BC = AD–|B|²,

    evidently a real number, which is called the discriminant . For a real circle as represented by the equation (1.1), Δ = – ρ². For the circle given by the equation (1.2), according to (1.3) the discriminant is found to be equal to

    (1.41)   Δ = –A²ρ².

    given by the equation (1.2) is an ordinary real circle if and only if A ≠ 0 and Δ < 0. Its centre γ and radius ρ can then be found from (1.3), (1.41), and we shall also denote it by (γ, ρ). It will have degenerated into a straight line if A = 0; in fact, in this case (1.2) turns out to be a linear equation in x and y with real coefficients. For any A ≠ 0 and Δ = 0 the circle appears as a point circle: ρ = 0.

    Finally we have the case Δ > 0. According to (1.41) this requires that ρnow is an ‘imaginary circle’ with a pure imaginary radius ρ and real centre γ. This circle may still be represented by the symbol (γ, ρ) and as in the case of a real circle γ and ρ may be found from (1.3) and (1.41). An imaginary circle has no real point, which means: no point represented by a complex number z = x + iy where x, y are real numbers. As an example we take the ‘imaginary unit circle’

    Obviously this equation cannot be solved by an ordinary complex number z. It may be written, however, in the form x² + y² = –1 and this equation can be solved by ‘points’ (x, y) whose coordinates x, y are not both real numbers,¹ for example, x = 0, y = i.

    Classification of Circles

    A ≠ 0: Δ < 0 real circle      ρ² > 0

    Δ = 0 point circle             ρ² = 0

    Δ > 0 imaginary circle     ρ² < 0.

    A = 0: Then Δ = –|B|² ≤ 0 always

    Δ < 0 straight line

    Δ = 0 no circle: B = C = 0 (cf. § 3, ex. 7).

    By (1.41)

    is said to have a positive sense of direction or positive orientation if A will be the same circle with negative orientation. The orientation of a directed circle may be indicated by an arrow on the circumference. Positive orientation corresponds to an anti-clockwise motion on the circumference leaving the interior of the circle at the left side.

    The notion of direction or orientation can be extended formally to imaginary circles, the sense of direction being determined by the sign of A.

    b.

    Two circles. 2 are not proportional (not linearly dependent) so that there is no real number λ 2=λ 1. We now study the one-parameter family of circles

    2,       λ1, λ2 real, not both zero.

    It is called a pencil of circles. The discriminant of this pencil is, by definition, the determinant

    (1.5)    

    that is, a quadratic form in the real variables λ1, λ2 with the real coefficients

    (1.51)    

    Now let A1A2 are the circles (γ1, ρ1), (γ2, ρ2) then by (1.3) and (1.41)

    (1.52)     Δ1=–A1²ρ1²,       Δ2=–A2²ρ2²,       2Δ12=A1A2(δ²–ρ²1–ρ²2)

    where

    δ = |γ1 – γ2|

    is the distance between their centres.

    2 are both real circles (Aj ≠ 0, Δj < 0, j = 1, 2) and that they have at least one real point in common (cf. Fig. 1). An orientation is fixed on these circles by the sign of the coefficients Aj. Then the angle ω 2 will be defined as the angle between the tangents at a common point taken in the direction defined by the orientation. By elementary geometry

    FIG. 1

    Hence

    j, and

    2 to have points in common it is necessary and sufficient that ω is a real angle:

    (1.62)   –1 ≤ cosω ≤ +1.

    By (1.6) this is equivalent to

    (1.63)   Δ1Δ2 – Δ²12 ≥ 0

    which together with Δ1 < 0 is the condition for the quadratic form (1.5) to have only non-positive values for all real λ1, λ2.

    The equality signs in (1.62) and (1.63) indicate contact or coincidence of the two circles. Suppose that both have positive orientation. If cos ω = +1, then ω = 0 and therefore δ² = (ρ1 – ρ2)² so that the smaller circle touches the greater one from inside. Similarly if cos ω = –1, then ω = π, δ = ρ1 + ρ2, and the two circles touch each other from outside.

    Evidently the right-hand term of (1.61) represents a real number for any pair of real or imaginary oriented (directed) circles. We call it their common invariant and denote it by cos ω, even though a real angle ω 2 this will be so if the quadratic form (1.5) may assume positive as well as negative values for different pairs λ1, λ2, that is, if

    (1.7)   Δ1Δ2 – Δ²12 < 0;

    then either

    cos ω > + 1,

    indicating that the smaller circle is entirely contained within the greater one, or

    cos ω < – 1,

    in which case the two circles are lying outside each other. This is readily seen if (1.6) is written in the form

    Indeed cos ω > 1 means |ρ1 – ρ2| > δ, and cos ω < – 1 similarly ρ1 + ρ2 < δ. The same follows by ‘continuity’ from the preceding discussion of the two different cases of contact.

    By (2, are orthogonal that is, perpendicular to each other, if and only if

    (1.72)   δ² = ρ²1 + ρ²2

    With regard to (1.61) the notion of orthogonality may be extended. A pair of circles, not necessarily real but not point circles (Δj ≠ 0), is said to be orthogonal if

    (1.73)   Δ12 = 0,

    which in the case of real circles is equivalent to (1.72).

    c. Pencils of circles. 2 are said to be generators 2. Any two different circles of the pencil may be taken as generators of the same pencil. Indeed, if with constant μj, νj

    are two circles of the pencil, they generate the pencil

    which coincides with the given pencil provided that for no real λ3,

    Enjoying the preview?
    Page 1 of 1