This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.Lecture 14 Poincaré Extension of Möbius Transformations
Given sphere preserving (Möbius) transformations in
n-dimensional
Euclidean space one can use the Poincaré extension to obtain
sphere preserving transformations in a half-space of n+1
dimensions. The Poincaré extension is usually provided either by
an explicit formula or by some geometric construction. We
investigate its algebraic background and describe all available
options. The solution is given in terms of one-parameter subgroups
of Möbius transformations acting on triples of quadratic forms.
To focus on the concepts, this paper deals with the Möbius
transformations of the real line only.
It is not surprising that we will arrive to the three possible situation—elliptic, parabolic, hyperbolic—considered above. The interesting feature of this chapter is what we completely avoid usage of complex, dual and double numbers which were are permanent companions so far.
14.1 Introduction
It is known, that Möbius transformations on ℝn can be
expanded to the “upper” half-space in ℝn+1 using the
Poincaré extension [24]*§ 3.3
[273]*§ 5.2. An explicit formula is usually
presented without a discussion of its origin. In particular, one may
get an impression that the solution is unique. This paper considers
various aspects of such extension and describes different possible
realisations. Our consideration is restricted to the case of extension
from the real line to the upper half-plane. However, we made an effort
to present it in a way, which allows numerous further generalisations.
14.2 Geometric construction
We start from the geometric procedure in the standard situation. The
group SL2(ℝ) consists of real 2× 2 matrices with the unit determinant.
SL2(ℝ) acts on the real line by linear-fractional maps:
| | : x ↦ | | ,
where
x∈ℝ and
| | ∈SL2(ℝ).
(1) |
A pair of (possibly equal) real numbers x
and y
uniquely determines a semicircle Cxy
in the upper half-plane with the diameter [x,y].
For a linear-fractional transformation
M (1),
the images M(x)
and M(y)
define the semicircle with the diameter [M(x),M(y)],
thus, we can define the action of M
on semicircles: M(Cxy)=CM(x)M(y).
Geometrically, the Poincaré extension
is based on the following
lemma, see Fig. 14.1(a) and more general
Lem. 18 below:
Lemma 1
If a pencil of semicircles in the upper half-plane has a common
point, then the images of these semicircles under a
transformation (1) have a common point as
well.
Elementary geometry of right triangles tells that a pair of
intersecting intervals [x,y], [x′,y′], where x<x′<y<y′,
defines the point
| ⎛
⎜
⎜
⎝ | | ,
| | ⎞
⎟
⎟
⎠ | ∈ ℝ+2.
(2) |
An alternative demonstration uses three observations:
- the scaling x↦ ax,
a>0
on the real line produces the scaling (u,v) ↦(au,av)
on pairs (2);
- the horizontal shift x↦ x+b on the real line produces
the horizontal shift (u,v) ↦ (u+b,v) on
pairs (2);
- for the special case y=−x−1 and y′=−x′−1
the pair (2) is (0,1).
Finally, expression (2), as well
as (3)–(4) below,
can be calculated by the specialised CAS for Möbius invariant
geometries [186, 211].
(a)
(b)
(c)
(d)
(e)
(f)
Figure 14.1: Poincaré extensions: first column
presents points defined by the intersecting intervals [x,y]
and [x′,y′], the second column—by disjoint intervals. Each
row uses the same type of conic sections—circles, parabolas and
hyperbolas respectively. Pictures are produced by the software [211]. |
This standard approach can be widened as follows. The above semicircle
can be equivalently described through the unique circle passing x
and y
and orthogonal to the real axis. Similarly, an interval [x,y]
uniquely defines a right-angle hyperbola in ℝ2
orthogonal to the real line and passing (actually, having her vertices
at) (x,0)
and (y,0).
An intersection with the second such hyperbola having vertices (x′,0)
and (y′,0)
defines a point with coordinates, see
Fig. 14.1(f):
where x<y<x′<y′. Note, the opposite sign of the product under the
square roots in (2)
and (3).
If we wish to consider the third type of conic
sections—parabolas—we cannot use the unaltered procedure: there is
no a non-degenerate parabola orthogonal to the real line and
intersecting the real line in two points. We may recall, that a circle
(or hyperbola) is orthogonal to the real line if its centre belongs to
the real line. Analogously, a parabola is focally orthogonal (see
[198]*§ 6.6 for a general consideration) to the real line
if its focus belongs to the real line. Then, an
interval [x,y] uniquely defines a downward-opened parabola with
the real roots x and y and focally orthogonal to the real line. Two
such parabolas defined by intervals [x,y] and [x′,y′] have a
common point, see Fig. 14.1(c):
| ⎛
⎜
⎜
⎝ |
| | ,
| (x′−x) (y′−y) (y−x+y′−x′)+(x+y−x′−y′) D |
|
(x−y−x′+y′)2 |
|
| ⎞
⎟
⎟
⎠ | ,
(4) |
where D=±√(x−x′)(y−y′)(y−x)(y′−x′).
For pencils of such hyperbolas and parabolas respective variants of
Lem. 1 hold.
Focally orthogonal parabolas make the angle 45∘ with the real
line. This suggests to replace orthogonal circles and hyperbolas by
conic sections with a fixed angle to the real line, see
Fig. 14.1(b)–(e). Of course, to be consistent
this procedure requires a suitable modification of
Lem. 1, we will obtain it as a byproduct
of our study, see Lem. 18. However, the
respective alterations of the above formulae
(2)–(4) become
more complicated in the general case.
The considered geometric construction is elementary and visually
appealing. Now we turn to respective algebraic consideration.
14.3 Möbius transformations and Cycles
The group SL2(ℝ) acts on ℝ2 by matrix multiplication on
column vectors:
Lg:
| |
↦
| | =
| | |
, where g= | | ∈SL2(ℝ).
(5) |
A linear action respects the equivalence relation (
)
∼ (
), λ≠ 0 on ℝ2. The
collection of all cosets for non-zero vectors in ℝ2 is the projective
line Pℝ1. Explicitly, a
non-zero vector (
)∈ℝ2 corresponds to the point with
homogeneous coordinates
[x1:x2]∈ Pℝ1. If x2≠ 0 then this point is
represented by [x1/x2:1] as well.
The embedding ℝ → Pℝ1 defined by x ↦
[x:1],
x∈ℝ covers the all but one
of points in Pℝ1. The exceptional point [1:0] is
naturally identified with the infinity.
The linear action (5) induces a morphism of
the projective line Pℝ1,
which is called a Möbius transformation. Considered on the real line
within Pℝ1,
Möbius transformations takes fraction-linear form:
g:
[x:1]
↦
| ⎡
⎢
⎢
⎣ | | :1 | ⎤
⎥
⎥
⎦ | , where g= | | ∈SL2(ℝ) and cx+d≠ 0.
|
This SL2(ℝ)-action on Pℝ1 is denoted as g: x↦
g· x. We note that the correspondence of
column vectors and row vectors
i: (
) ↦ (x2, −x1) intertwines
the left multiplication Lg (5) and
the right multiplication Rg−1 by the inverse matrix:
Rg−1:
(x2,−x1)
↦
(cx1+dx2, −ax1−bx2)
=
(x2,−x1)
| | .
(6) |
We extended the map i to 2× 2-matrices by the rule:
Two columns (
) and (
) form the 2× 2 matrix Mxy=(
). For geometrical reasons appearing in
Cor. 3, we call a cycle the
2× 2-matrix Cxy defined by
Cxy= | | Mxy·i(Mxy)= | | Myx·i(Myx)=
| ⎛
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎝ | | ⎞
⎟
⎟
⎟
⎟
⎟
⎟
⎟
⎠ |
| .
(8) |
We note that detCxy=−(x−y)2/4, thus detCxy=0 if and
only if x=y. Also, we can consider the Möbius transformation
produced by the 2× 2-matrix Cxy and calculate:
Cxy
| | =
λ | |
and
Cxy
| | =
−λ | |
where λ= | | .
(9) |
Thus, points [x:1], [y:1]∈ Pℝ1 are fixed by
Cxy. Also, Cxy swaps the interval [x,y] and its
complement.
Due to their structure, matrices Cxy can be parametrised by
points of ℝ3. Furthermore, the map from
ℝ2→ ℝ3 given by (x,y)↦
Cxy naturally induces the projective map (Pℝ1)2
→ Pℝ2 due to the identity:
| | |
| | =λ µ | ⎛
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎝ | | ⎞
⎟
⎟
⎟
⎟
⎟
⎟
⎟
⎠ |
| = λ µ Cxy .
|
Conversely, a zero-trace matrix
(
)
with a non-positive determinant is projectively equivalent to
a product Cxy (8) with
x,y=a±√a2+bc/c. In particular, we can embed a point
[x:1]∈ Pℝ1 to 2× 2-matrix Cxx with
zero determinant.
The combination
of (5)–(8) implies
that the correspondence (x,y)↦ Cxy is SL2(ℝ)-covariant in the
following sense:
g Cxy g−1=Cx′y′ , where
x′=g· x and y′=g· y.
(10) |
To achieve a geometric interpretation of all matrices, we consider the
bilinear form Q: ℝ2× ℝ2→
ℝ generated by a 2× 2-matrix (
):
Q(x, y)=
| |
| |
| | , where x=(x1,x2), y=(y1,y2).
(11) |
Due to linearity of Q, the null set
{(x,y)∈ ℝ2× ℝ2 ∣ Q(x,y)=0}
(12) |
factors to a subset of Pℝ1×
Pℝ1. Furthermore, for the matrices
Cxy (8), a direct calculation
shows that:
Lemma 2
The following identity holds:
Cxy(i(x′),y′)=tr(Cxy Cx′y′)
= | | (x+y)(x′+ y′)−(x y+x′ y′) .
(13) |
In particular, the above expression is a symmetric function of the pairs
(
x,
y)
and (
x′,
y′)
.
The map i appearance in (13) is
justified once more by the following result.
Corollary 3
The null set of the quadratic form
Cxy(
x′)=
Cxy(
i(
x′),
x′)
consists of two points x
and y.
Alternatively, the identities Cxy(x)=Cxy(y)=0 follows
from (9) and the fact that i(z) is
orthogonal to z for all z∈ℝ2. Also, we note that:
Motivated by Lem. 2, we call
⟨ Cxy,Cx′y′
⟩:=−tr(Cxy Cx′y′) the
pairing of two cycles. It shall be noted that
the pairing is not positively defined, this follows from the
explicit expression (13). The sign is chosen
in such way, that
⟨ Cxy,Cxy
⟩=−2det(Cxy)= | | (x−y)2≥ 0.
|
Also, an immediate consequence of Lem. 2 or
identity (11) is
Corollary 4
The pairing of cycles is invariant under the
action (10) of SL2(ℝ)
:
⟨ g· Cxy · g−1 ,g· Cx′y′ · g−1
⟩ =⟨ Cxy,Cx′y′
⟩.
|
From (13), the null
set (12) of the form Q=Cxy can be
associated to the family of cycles {Cx′y′ ∣
⟨ Cxy,Cx′y′
⟩=0, (x′,y′)∈ℝ2×
ℝ2} which we will call orthogonal to Cxy.
14.4 Extending cycles
Since bilinear forms with matrices Cxy have numerous geometric
connections with Pℝ1, we are looking for a similar
interpretation of the generic matrices. The previous discussion
identified the key ingredient of the recipe: SL2(ℝ)-invariant
pairing (13) of two forms. Keeping in mind
the structure of Cxy, we will parameterise1 a
generic 2× 2 matrix as (
) and consider the corresponding four dimensional vector
(n,l,k,m). Then, the similarity with (
)
∈SL2(ℝ):
corresponds to the linear transformation of ℝ4,
cf. [198]*Ex. 4.15:
| | =
| ⎛
⎜
⎜
⎜
⎝ | 1 | 0 | 0 | 0 |
0 | c b+a d | b d | c a |
0 | 2 c d | d2 | c2 |
0 | 2 a b | b2 | a2
|
| ⎞
⎟
⎟
⎟
⎠ |
|
| | .
(15) |
In particular, this action commutes with the scaling of the first component:
λ: (n,l,k,m) ↦ (λ n,l,k,m).
(16) |
This expression is helpful in proving the following statement.
Lemma 5
Any SL2(ℝ)
-invariant (in the sense of the
action (15)) pairing in ℝ
4 is
isomorphic to
where σ
c =−1
, 0
or 1
and
(
n,
l,
k,
m)
, (
n′,
l′,
k′,
m′)∈ℝ
4 .
Proof.
Let T be 4× 4 a matrix
from (15), if a SL2(ℝ)-invariant pairing
is defined by a 4× 4 matrix J=(jfg), then
T′JT=J, where T′ is transpose of T. The equivalent
identity T′J=JT−1 produces a system of homogeneous linear
equations which has the generic solution:
| j12 | =
j13=
j14=
j21=
j31=
j41=0, | | | | | | | | | |
j22 | |
j23 | |
j24 | | | | | | |
j34 | = | c (a − d) j42+(a −d)2j44 |
|
c2 |
| + j43, |
|
j33 | |
j32 | | | | | | |
|
with four free variables j11, j42, j43 and
j44. Since a solution shall not depend on a, b,
c, d, we have to put j42=j44=0. Then by the
homogeneity of the identity T′J=JT−1, we can scale j43
to 1. Thereafter, an independent (sign-preserving)
scaling (16) of n leaves only three
non-isomorphic values −1, 0, 1 of j11.
The appearance of the three essential different cases σc =−1,
0 or 1 in Lem. 5 is a
manifestation of the common division of mathematical objects into
elliptic, parabolic and hyperbolic cases [185]
[198]*Ch. 1. Thus, we will use letters “e”, “p”,
“h” to encode the corresponding three values of σc.
Now we may describe all SL2(ℝ)-invariant pairings of bilinear forms.
Corollary 6
Any SL2(ℝ)
-invariant (in the sense of the
similarity (14)) pairing between
two bilinear forms Q= (
)
and Q′= (
)
is isomorphic to:
|
⟨ Q,Q′
⟩τ | =−tr(QτQ′) | | | | | | | | | (17) |
| =
2τ n′n−2l′l+k′m+m′k, where
Qτ= | |
| | | | | | | | | |
|
and τ=−1
, 0
or 1
.
Note that we can explicitly write Qτ for Q=
(
) as follows:
Qe= | | ,
Qp= | ⎛
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎝ | | ⎞
⎟
⎟
⎟
⎟
⎟
⎟
⎟
⎠ |
| ,
Qh= | | .
|
In particular, Qh=−Q−1 and
Qp=1/2(Qe+Qh). Furthermore, Qp has the same
structure as Cxy. Now, we are ready to extend the
projective line Pℝ1 to two dimensions using the analogy
with properties of cycles Cxy.
Definition 7
-
Two bilinear forms Q and Q′ are
τ-orthogonal if ⟨ Q,Q′
⟩τ=0.
-
A form is τ-isotropic if it is
τ-orthogonal to itself.
If a form Q= (
)
has k≠ 0 then we can scale it to obtain k=1, this form
of Q
is called normalised. A normalised τ-isotropic form is
completely determined by its diagonal values: (
). Thus, the set of such forms is in a one-to-one
correspondence with points of ℝ2. Finally, a form
Q= (
) is e-orthogonal to the τ-isotropic form (
) if:
k(u2−τ v2)−2lu−2nv+m=0,
(18) |
that is the point (u,v)∈ℝ2 belongs to the quadratic curve
with coefficients (k,l,n,m).
14.5 Homogeneous spaces of cycles
Obviously, the group SL2(ℝ) acts on Pℝ1 transitively,
in fact it is even 3-transitive in the following sense. We say
that a triple {x1,x2,x3}⊂ Pℝ1 of distinct
points is positively oriented if
either x1<x2<x3, or x3<x1< x2,
|
where we agree that the ideal point ∞∈ Pℝ1 is
greater than any x∈ℝ. Equivalently,
a triple {x1,x2,x3} of reals is positively oriented if:
Also, a triple of distinct points, which is not positively oriented, is
negatively oriented.
A simple calculation based on the resolvent-type identity:
| | − | | = | (x−y)(ad−bc) |
|
(cx+b)(cy+d) |
|
shows that both the positive and negative orientations of triples are
SL2(ℝ)-invariant. On the other hand, the reflection x↦ −x
swaps orientations of triples. Note, that the reflection is a Moebius
transformation associated to the cycle
C0∞= | | , with detC0∞=−1.
(19) |
A significant amount of information about Möbius transformations
follows from the fact, that any continuous one-parametric subgroup of
SL2(ℝ) is conjugated to one of the three following
subgroups2:
A= | ⎧
⎨
⎩ | | ⎫
⎬
⎭ | ,
N= | ⎧
⎨
⎩ | | ⎫
⎬
⎭ | ,
K= | ⎧
⎨
⎩ | | ⎫
⎬
⎭ | ,
(20) |
where t∈ℝ. Also, it is useful to introduce
subgroups and N′ conjugated to A and N
respectively:
= | ⎧
⎨
⎩ | | ∣ t∈ ℝ | ⎫
⎬
⎭ | ,
N′= | ⎧
⎨
⎩ | | ∣ t∈ ℝ | ⎫
⎬
⎭ | .
(21) |
Thereafter, all three one-parameter subgroups , N′
and K consist of all matrices with the universal structure
| |
where τ =1, 0, −1 for , N′
and
K respectively.
(22) |
We use the notation Hτ for these subgroups. Again, any
continuous one-dimensional
subgroup of SL2(ℝ) is conjugated to
Hτ for an appropriate τ.
We note, that matrices from A, N and K with t≠ 0
have two, one and none different real eigenvalues
respectively. Eigenvectors in ℝ2 correspond to fixed
points of Möbius transformations on Pℝ1. Clearly, the
number of eigenvectors (and thus fixed points) is limited by the
dimensionality of the space, that is two. For this reason, if g1
and g2 take equal values on three different points of
Pℝ1, then g1=g2.
Also, eigenvectors provide an effective classification tool:
g∈SL2(ℝ) belongs to a one-dimensional continuous subgroup conjugated
to A, N or K if and only if the characteristic polynomial
det(g−λ I) has two, one and none different real root(s)
respectively. We will illustrate an application of fixed points
techniques through the following well-known result, which will be used
later.
Lemma 8
Let {
x1,
x2,
x3}
and {
y1,
y2,
y3}
be positively
oriented triples of points in ℂ
. Then, there is a
unique (computable!) Möbius map φ∈
SL2(ℝ)
with φ(
xj)=
yj for
j=1
, 2
, 3
.
Proof.
Often, the statement is quickly demonstrated through an explicit
expression for φ, cf. [26]*Thm. 13.2.1. We
will use properties of the subgroups A, N and K to
describe an algorithm to find such a map. First, we notice that it
is sufficient to show the Lemma for the particular case y1=0,
y2=1, y3=∞. The general case can be obtained from
composition of two such maps. Another useful observation is that the
fixed point for N, that is ∞, is also a fixed point of
A.
Now, we will use subgroups K, N and A in order of
increasing number of their fixed points. First, for any x3 the
matrix g′=(
)∈ K such that cott=x3 maps x3 to
y3=∞. Let x1′=g′x1 and x2′=g′x2. Then the matrix
g″=(
)∈ N, fixes ∞=g′ x3 and sends x′1 to
y1=0. Let x″2=g″x2′, from positive orientation of
triples we have 0<x″2<∞. Next, the matrix
g‴=(
)∈ A with a=√x″2 sends x″2 to 1
and fixes both
∞=g″g′x3 and 0=g″g′ x1. Thus, g=g‴g″g′ makes
the required transformation (x1,x2,x3)↦(0,1,∞).
Corollary 9
Let {x1,x2,x3} and {y1,y2,y3} be two triples with the
opposite orientations. Then, there is a
unique Möbius map φ∈SL2(ℝ) with φ∘ C0∞(xj)=yj for
j=1, 2, 3.
We will denote by φXY the unique map from
Lem. 8 defined by triples
X={x1,x2,x3} and Y={y1,y2,y3}.
Although we are not going to use it in this paper, we note that the
following important result [240]*§ III.1 is an immediate
consequence of our proof of Lem. 8.
Corollary 10 (Iwasawa decomposition)
Any element of g∈
SL2(ℝ)
is a product g=
gA gN gK, where
gA, gN and gK belong to subgroups A, N,
K respectively and those factors are uniquely defined.
In particular, we note that it is not a coincidence that the subgroups
appear in the Iwasawa decomposition SL2(ℝ)=ANK in order of decreasing
number of their fixed points.
14.6 Triples of intervals
We change our point of view and instead of two ordered triples of points
consider three ordered pairs, that is—three intervals. For them we
will need the following definition.
Definition 11
We say that a triple of intervals {[
x1,
y1], [
x2,
y2],
[
x3,
y3]}
is aligned
if the triples X={
x1,
x2,
x3}
and Y={
y1,
y2,
y3}
of their endpoints have the same orientation.
Aligned triples determine certain one-parameter subgroups of Möbius
transformations as follows:
Proposition 12
Let {[
x1,
y1], [
x2,
y2], [
x3,
y3]}
be an aligned triple of
intervals.
-
If φXY has at most one fixed point, then there is a
unique (up to a parametrisation)
one-parameter semigroup of
Möbius map ψ(t)⊂SL2(ℝ), which maps [x1,y1] to
[x2,y2] and [x3,y3]:
ψ(tj)(x1)=xj, ψ(tj)(y1)=yj, for
some tj∈ℝ and j=2,3.
|
- Let φXY have two fixed points x<y and Cxy
be the orientation inverting Möbius transformation with the
matrix (8). For j=1,
2, 3, we define:
| x′j | =xj, | y′j | =yj, | x″j | =Cxyxj, |
y″j | =Cxyyj | if x<xj<y; | | |
x′j | =Cxyxj, | y′j | =Cxyyj, | x″j | =xj, | y″j | =yj, |
otherwise.
| | |
|
Then, there is a one-parameter
semigroup of Möbius map ψ(t)⊂SL2(ℝ), and t2,
t3∈ℝ such that:
ψ(tj)(x′1)=x′j, ψ(tj)(x″1)=x″j,
ψ(tj)(y′1)=y′j, ψ(tj)(y″1)=y″j,
|
where j=2,3.
Proof.
Consider the one-parameter subgroup of
ψ(t)⊂ SL2(ℝ) such that φXY=ψ(1). Note, that
ψ(t) and φXY have the same fixed points (if any) and no
point xj is fixed since xj≠ yj. If the number of fixed
points is less than 2, then ψ(t)x1, t∈ℝ
produces the entire real line except a possible single fixed
point. Therefore, there are t2 and t3 such that
ψ(t2)x1=x2 and ψ(t3)x1=x3. Since ψ(t) and
φXY commute for all t we also have:
ψ(tj)y1=ψ(tj)φXY x1=φXYψ(tj) x1=φXY
xj=yj, for j=2,3.
|
If there are two fixed points x<y, then the open interval
(x,y) is an orbit for the subgroup ψ(t). Since all
x′1, x′2 and x′3 belong to this orbit and Cxy
commutes with φXY we may repeat the above reasoning for
the dashed intervals [x′j,y′j]. Finally, x″j=Cxyx′j
and y″j=Cxyy′j, where Cxy commutes with φ
and ψ(tj), j=2, 3. Uniqueness of the subgroup
follows from Lemma 13.
The group SL2(ℝ) acts transitively on collection of all cycles
Cxy, thus this is a SL2(ℝ)-homogeneous space. It is easy to
see that the fix-group of the cycle C−1,1 is
(21). Thus the homogeneous space of
cycles is isomorphic to SL2(ℝ)/.
Lemma 13
Let H be a one-parameter continuous subgroup of SL2(ℝ)
and
X=
SL2(ℝ)/
H be the corresponding homogeneous space. If two orbits
of one-parameter continuous subgroups on X have at least three
common points then these orbits coincide.
Proof.
Since H is conjugated either to , N′ or K,
the homogeneous space X=SL2(ℝ)/H is isomorphic to the upper
half-plane in double, dual or complex
numbers [198]*§ 3.3.4. Orbits of one-parameter
continuous subgroups in X are conic sections, which are circles,
parabolas (with vertical axis) or equilateral hyperbolas (with
vertical axis) for the respective type of geometry. Any two
different orbits of the same type intersect at most at two points,
since an analytic solution reduces to a quadratic equation.
Alternatively, we can reformulate Prop. 12 as
follows: three different cycles Cx1y1, Cx2y2,
Cx3y3 define a one-parameter subgroup, which generate either one
orbit or two related orbits passing the three cycles.
We have seen that the number of fixed points is the key
characteristics for the map φXY. The next result gives an
explicit expression for it.
Proposition 14
The map φ
XY has zero, one or two fixed points
if the expression
det
| ⎛
⎜
⎜
⎝ | 1 | x1y1 | y1−x1 |
1 | x2y2 | y2−x2 |
1 | x3y3 | y3−x3
|
| ⎞
⎟
⎟
⎠ | |
| −4det | |
·det
| ⎛
⎜
⎜
⎝ | x1 | −x1y1 | y1 |
x2 | −x2y2 | y2 |
x3 | −x3y3 | y3
|
| ⎞
⎟
⎟
⎠ |
|
(23) |
is negative, zero or positive respectively.
Proof.
If a Möbius transformation (
) maps x1↦ y1, x2↦ y2,
x3↦ y3 and s↦ s, then we have a homogeneous linear
system, cf. [26]*Ex. 13.2.4:
| ⎛
⎜
⎜
⎜
⎝ | x1 | 1 | −x1y1 | −y1 |
x2 | 1 | −x2y2 | −y2 |
x3 | 1 | −x3y3 | −y3 |
s | 1 | −s2 | −s
|
| ⎞
⎟
⎟
⎟
⎠ |
|
| | =
| | .
(24) |
A non-zero solution exists if the determinant of the 4× 4
matrix is zero. Expanding it over the last row and
rearranging terms we obtain the quadratic equation for the fixed point
s:
s2 det
| |
+sdet
| ⎛
⎜
⎜
⎝ | 1 | x1y1 | y1−x1 |
1 | x2y2 | y2−x2 |
1 | x3y3 | y3−x3
|
| ⎞
⎟
⎟
⎠ |
|
+det
| ⎛
⎜
⎜
⎝ | x1 | −x1y1 | y1 |
x2 | −x2y2 | y2 |
x3 | −x3y3 | y3
|
| ⎞
⎟
⎟
⎠ |
|
=0.
|
The value (23) is the discriminant of this equation.
Remark 15
It is interesting to note, that the relation ax+
b−
cxy−
dy=0
used
in (24)
can be stated as e-orthogonality of the cycle
(
)
and the isotropic bilinear form
(
)
.
If y=g0· x for some g0∈
Hτ, then for any g∈ Hτ we also have
yg=g0· xg, where xg=g· xg and yg=g·
yg. Thus, we demonstrated the first part of the following result.
Lemma 16
Let τ =1
, 0
or −1
and a real constant t≠ 0
be such that
1−τ
t2>0
.
-
The collections of intervals:
Iτ,t = | ⎧
⎪
⎨
⎪
⎩ | [x, | | ] ∣ x∈ ℝ | ⎫
⎪
⎬
⎪
⎭ | (25) |
is preserved by the actions of subgroup Hτ. Any three
different intervals from Iτ,t define the subgroup
Hτ in the sense of Prop. 12.
- All Hτ-invariant bilinear forms compose the family Pτ,t={()
}.
The family Pτ,t consists of the eigenvectors of the
4× 4 matrix from (15) with
suitably substituted entries. There is (up to a factor) exactly one
τ-isotropic form in Pτ,t, namely (
). We denote this form by ι. It corresponds to
the point (0,1)∈ℝ2 as discussed after
Defn. 7. We may say that the subgroup
Hτ fixes the point ι, this will play an important
rôle below.
14.7 Geometrisation of cycles
We return to the geometric version of the Poincaré extension
considered in Sec. 14.2 in terms of cycles.
Cycles of the form (
) are τ-isotropic for any τ and are
parametrised by the point x of the real line. For a fixed
τ, the collection of all τ-isotropic cycles is a larger
set containing the image of the real line from the previous
sentence. Geometrisation of this embedding is described in the
following result.
Lemma 17
-
The transformation x↦ x+τ t/tx+1 from the
subgroup Hτ, which maps x↦ y, corresponds to
the value t=x−y/xy−τ.
- The unique (up to a factor) bilinear form Q orthogonal
to Cxx, Cyy and ι is
Q= | ⎛
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎝ | | ⎞
⎟
⎟
⎟
⎟
⎟
⎟
⎟
⎠ |
| .
|
- The defined above t and Q are connected by the cycles cross ratio (15) identity:
Here, the real line is represented by the bilinear form ℝ=
()
defined by the identity matrix.
- For a cycle Q=
()
, the value of the cycles cross ratio
is equal to the square of cosine of the angle between the curve k(u2+τ
v2)−2lu−2nv+m=0 (18) and the real line,
cf. Ex. 1 and Ex 8.
Proof.
The first statement is verified by a short calculation. A form Q=
(
)
in the
second statement may be calculated from the homogeneous system:
which has the rank 3 if x≠ y. The third statement
can be checked by a calculation as well. Finally, the last item is a
particular case of the more general statement as indicated. Yet, we
can derive it here from the implicit derivative
dv/du=ku−l/n of the function k(u2+τ
v2)−2lu−2nv+m=0 (18) at the point
(u,0). Note that this value is independent from τ. Since
this is the tangent of the intersection angle with the real line,
the square of the cosine of this angle is:
if ku2−2 u l+m=0.
Also, we note that, the independence of the left-hand side
of (26) from x can be shown
from basic principles. Indeed, for a fixed t the subgroup
Hτ acts transitively on the family of triples {x,
x+τ t/tx+1, ι}, thus Hτ acts transitively on
all bilinear forms orthogonal to such triples. However, the left-hand
side of (26) is SL2(ℝ)-invariant, thus may not
depend on x. This simple reasoning cannot provide the exact
expression in the right-hand side of (26),
which is essential for the geometric interpretation of the Poincaré
extension.
To restore a cycle from its intersection points with the real line we
need also to know its cycle product with the real line. If this
product is non-zero then the sign of the parameter n is
additionally required. At the cycles’ language, a common point of
cycles C and S
is encoded by a cycle
Ĉ such that:
⟨ G ,C
⟩e=
⟨ G ,S
⟩e=
⟨ G ,G
⟩τ=0.
(27) |
For a given value of τ, this produces two linear and one
quadratic equation for parameters of Ĉ. Thus, a pair of
cycles may not have a common point or have up to two such
points. Furthermore, Möbius-invariance of the above
conditions (26)
and (27) supports the geometrical
construction of Poincaré extension, cf. Lem. 1:
Lemma 18
Let a family consist of cycles, which are e-orthogonal to a
given τ
-isotropic cycle G and have the
fixed value of the fraction in the left-hand side
of (26). Then, for a given Möbius
transformation g and any cycle C from the
family, gC is e-orthogonal to the
τ
-isotropic cycle gG and has the same
fixed value of the fraction in the left-hand side
of (26) as C.
Summarising the geometrical construction, the Poincaré extension based on
two intervals and the additional data produces two situations:
- If the cycles C and C′ are orthogonal to the real line, then
a pair of overlapping cycles produces a point of the elliptic upper
half-plane, a pair of disjoint cycles defines a point of the
hyperbolic. However, there is no orthogonal cycles uniquely defining
a parabolic extension.
- If we admit cycles, which are not orthogonal to the real line,
then the same pair of cycles may define any of the three different
types (EPH) of extension.
These peculiarities make the extension based on three intervals,
described above, a bit more preferable.
14.8 Concluding remarks
Based on the consideration in
Sections 14.3– 14.7 we describe the
following steps to carry out the generalised extension procedure:
- Points of the extended space are equivalence classes of aligned
triples of cycles in Pℝ1, see
Defn. 11. The equivalence relation
between triples will emerge at step 3.
- A triple T of different cycles
defines the unique one-parameter continues subgroup S(t) of
Möbius transformations as defined in
Prop. 12.
- Two triples of cycles are
equivalent if and only if the subgroups defined in
step 2 coincide (up to a parametrisation).
- The geometry of the extended space, defined by the equivalence
class of a triple T, is elliptic, parabolic or hyperbolic
depending on the subgroup S(t) being similar S(t)=gHτ(t)
g−1, g∈SL2(ℝ) (up to parametrisation) to
Hτ (22) with τ=−1, 0 or
1 respectively. The value of τ may be identified from the
triple using Prop. 14.
- For the above τ and g∈SL2(ℝ), the point of the
extended space, defined by the the equivalence class of a triple
T, is represented by τ-isotropic (see
Defn. 7(2)) bilinear form
g−1
()
g, which is S-invariant, see the end of
Section 14.6.
Obviously, the above procedure is more complicated that the geometric
construction from Section 14.2. There are reasons for
this, as discussed in Section 14.7: our procedure is
uniform and we are avoiding consideration of numerous subcases created
by an incompatible selection of parameters. Furthermore, our presentation
is aimed for generalisations to Möbius transformations of moduli
over other rings. This can be considered as an analog of Cayley–Klein
geometries [339]*Apps. A–B [284].
It shall be rather straightforward to adopt the extension for
ℝn. Möbius transformations in ℝn are
naturally expressed as linear-fractional transformations in Clifford
algebras [65] with a similar classification of subgroups
based on fixed points [2, 343]. The Möbius
invariant matrix presentation of cycles ℝn is already
known [65]*(4.12) [100]
[208]*§ 5. Of course, it is necessary to
enlarge the number of defining cycles from 3 to, say,
n+2. This shall have a connection with Cauchy–Kovalevskaya
extension considered in Clifford analysis [295, 310].
Naturally, a consideration of other moduli and rings may require some
more serious adjustments in our scheme.
Our construction is based on the matrix presentations of cycles. This
techniques is effective in many different cases
[198, 208]. Thus, it is not surprising that such ideas
(with some technical variation) appeared independently on many
occasions [65]*(4.12)
[100] [300]*§ 1.1
[163]*§ 4.2. The interesting feature of the present
work is the complete absence of any (hyper)complex numbers. It deemed
to be unavoidable [198]*§ 3.3.4 to employ complex, dual
and double numbers to represent three different types of Möbius
transformations extended from the real line to a plane. Also (hyper)complex
numbers were essential in [198, 185] to define
three possible types of cycle product (4),
and now we managed without them.
Apart from having real entries, our matrices for cycles share the
structure of matrices from [65]*(4.12)
[100] [198] [185]. To
obtain another variant, one replaces the map
i (7) by
Then, we may define symmetric matrices in a manner similar
to (8):
Cxyt= | | Mxy·t(Mxy)=
| ⎛
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎝ | | ⎞
⎟
⎟
⎟
⎟
⎟
⎟
⎟
⎠ |
| .
|
This is the form of matrices for cycles similar
to [300]*§ 1.1
[163]*§ 4.2. The property (10)
with matrix similarity shall be replaced by the respective one with
matrix congruence: g· Cxyt · gt=
Cx′y′t. The rest of our construction may be adjusted
for these matrices accordingly.
Last modified: October 28, 2024.