Real projective line

From Wikipedia, the free encyclopedia
Jump to: navigation, search
The real projective line is the real line together with a point at infinity; i.e., the one-point compactification of R.

In geometry, a real projective line is an extension of the usual concept of line that has been historically introduced to solve a problem set by visual perspective: two parallel lines do not intersect but seem to intersect "at infinity". For solving this problem, points at infinity have been introduced, in such a way that in a real projective plane, two different projective lines meet in exactly one point. The set of these points at infinity, the "horizon" of the visual perspective in the plane, is a real projective line. It is the circle of directions emanating from an observer situated at any point, with opposite points identified. A model of the real projective line is the projectively extended real line. Drawing a line to represent the horizon in visual perspective, an additional point at infinity is added to represent the collection of lines parallel to the horizon.

Formally, the real projective line is defined as the space of all one-dimensional linear subspaces of a two-dimensional vector space over the reals. Accordingly, the general linear group of 2×2 invertible matrices acts on the real projective line. Since the center acts trivially, the projective linear group, PGL(2, R), also acts on the projective line. These are the geometric transformations of the projective line. When the projective line is represented as a real line with point at infinity, the elements of the projective linear group act as fractional linear transformations. These transformations of the real projective line are called homographies.

Topologically, the real projective line is homeomorphic to the circle. The real projective line is the boundary of the hyperbolic plane. Every isometry of the hyperbolic plane induces a unique geometric transformation of the boundary, and vice versa. Furthermore, every harmonic function on the hyperbolic plane is given as a Poisson integral of a distribution on the projective line, in a manner that is compatible with the action of the isometry group. The topological circle has many compatible projective structures on it; the space of such structures is the (infinite dimensional) universal Teichmüller space. The complex analog of the real projective line is the complex projective line; that is, the Riemann sphere.


The points of the real projective line are usually defined as equivalence classes of an equivalence relation. The starting point is a real vector space of dimension 2, V. Define on V ∖ 0 the binary relation v ~ w to hold when there exists a nonzero real number t such that v = tw. The definition of a vector space implies almost immediately that this is an equivalence relation. The equivalence classes are the vector lines from which the zero vector has been removed. The real projective line P(V) is the set of all equivalence classes. Each equivalence class is considered as a single point, or, in other words, a point is defined as being an equivalence class.

If one chooses a basis of V, this amounts (by identifying a vector with its coordinates vector) to identify V with the direct product R × R = R2, and the equivalence relation becomes (x, y) ~ (w, z) if there exists a nonzero real number t such that (x, y) = (tw, tz). In this case, the projective line P(R2) is preferably denoted P1(R) or  \mathbb{R}\mathbb{P}^1. The equivalence class of the pair (x, y) is traditionally denoted [x: y], the colon in the notation recalling that, if y ≠ 0, the ratio x : y is the same for all elements of the equivalence class. If a point P is the equivalence class [x: y] one says that (x, y) is a pair of projective coordinates of P.[1]

As P(V) is defined through an equivalence relation, the canonical projection from V to P(V) defines a topology (the quotient topology) and a differential structure on the projective line. However, the fact that equivalence classes are not finite induces some difficulties for defining the differential structure. These are solved by considering V as an Euclidean vector space. The circle of the unit vectors is, in the case of R2, the set of the vectors whose coordinates satisfy x2 + y2 = 1. This circle intersects each equivalence classes in exactly two opposite points. Therefore, the projective line may be considered a the quotient space of the circle by the equivalence relation such that v ~ w if and only if either v = w or v = −w.


The projective line is a manifold. This can be seen by above construction through an equivalence relation, but is easier to understand by provideing an atlas consisting of two charts

  • Chart #1: y\ne 0, \quad [x: y] \mapsto \frac {x}{y}
  • Chart #2: x\ne 0, \quad [x: y] \mapsto \frac {y}{x}

The equivalence relation provides that all representatives of an equivalence class are sent to the same real number by a chart.

Either of x or y may be zero, but not both, so both charts are needed to cover the projective line. The transition map between these two charts is the multiplicative inverse. As it is an differentiable function, and even an analytic function (outside of zero), the real projective line is both a differentiable manifold and an analytic manifold.

The inverse function of chart #1 is the map

 x \mapsto [x: 1].

It defines an embedding of the real line into the projective line, whose complement of the image is the point [1: 0]. The pair consisting of this embedding and the projective line is called the projectively extended real line. Identifying the real line with its image by this embedding, one sees that the projective line may be considered as the union of the real line and the single point [1: 0], called the point at infinity of the projectively extended real line, and denoted . This embedding allows us to identify the point [x: y] either with the real number x/y if y ≠ 0, or with in the other case.

The same construction may be done with the other chart. In this case, the point at infinity is [0: 1]. This shows that the notion of point at infinity is not intrinsic to the real projective line, but is relative to the choice of an embedding of the real line into the projective line.


The real projective line is a complete projective range that is found in the real projective plane and in the complex projective line. Its structure is thus inherited from these superstructures. Primary among these structures is the relation of projective harmonic conjugates among the points of the projective range.

The real projective line has a cyclic order which is an important mathematical structure in showing that the real line is totally ordered and complete.[2] The cyclic order is addressed by a separation relation which has the properties necessary for appropriate deductions.


The mappings of P1(R) are called homographies or projectivities. These automorphisms can be constructed synthetically as central projections or parallel projections and their compositions. In homogeneous coordinates, automorphisms are given by the projective linear group PGL(2, R), which consists of all invertible 2 × 2 real matrices with proportional matrices identified.

Elements of PGL(2, R) can be realized concretely as fractional linear transformations of the form

x\mapsto \frac{ax+b}{cx+d}

where ad-bc\not=0, and x is an affine coordinate on the projective line.

The group PGL(2, R) is triply transitive on the real projective line, meaning that for any two triples of distinct points, there is a unique automorphism that maps the first triple onto the second. The stabilizer of any point (a "point at infinity") is the affine group of a line.

Because PGL(2, R) is isomorphic to the pseudo-orthogonal group SOo(1,2), a subgroup of the Lorentz group, it is possible to represent Lorentz transformations as automorphisms of the real projective line. For example, the Lorentz boost

f(x)=\frac {x + v/c}{xv/c + 1}

has the properties that f(1)=1 and f(–1) = –1 while f(0) = v/c.[3]

Since ℤ ⊂ ℝ ⊂ ℂ, the automorphism group PGL(2, R) lies between the modular group PGL(2, Z) and the Möbius group PGL(2, C).


  1. ^ The argument used to construct P1(R) can also be used with any field K and any dimension to construct the projective space Pn(K).
  2. ^ Bruce E. Meserve (1955) Fundamental Concepts of Geometry, p. 89, at Google Books
  3. ^ V.V. Prasolov & V.M. Tikhomirov, O.V. Sipacheva translator (2001) Geometry, pages 90, 138, 139, Translations of Mathematical Monographs 200, American Mathematical Society ISBN 0-8218-2038-9