M.E. Irizarry-Gelpí

Physics impostor. Mathematics interloper. Husband. Father.

Perplex Numbers as Pairs


A perplex number \(z\) can be written as

$$ z = a + s b; $$

where both \(a\) and \(b\) are real numbers, and \(s^{2} = +1\). Note that \(s \neq \pm 1\): \(s\) is a non-trivial element that squares to 1. Perplex numbers are also known as split-complex numbers. This post on perplex numbers follows closely the format of a previous post on complex numbers.

Another way to represent a perplex number is as a pair of real numbers:

$$ z = \begin{pmatrix} a & b \end{pmatrix}. $$

I will prefer this way since it is more compatible with how a perplex number is represented computationally. The left component is called the real part:

$$ \operatorname{Real}{(z)} = a; $$

and the right component is called the unreal part:

$$ \operatorname{Unreal}{(z)} = b. $$

Addition and Scaling

You can define two operations on perplex numbers. The addition operation is given by

$$ z_{1} = \begin{pmatrix} a_{1} & b_{1} \end{pmatrix}, \qquad z_{2} = \begin{pmatrix} a_{2} & b_{2} \end{pmatrix} \quad \longrightarrow \quad \operatorname{Add}{(z_{1}, z_{2})} = \begin{pmatrix} a_{1} + a_{2} & b_{1} + b_{2} \end{pmatrix}. $$

The scale operation is given by

$$ z = \begin{pmatrix} a & b \end{pmatrix} \quad \longrightarrow \quad \operatorname{Scale}{(z, c)} = \begin{pmatrix} c a & c b \end{pmatrix}. $$

The addition and scale operations are essential for vector spaces. Indeed, you can assign a perplex number to a 2-dimensional real vector and vice-versa:

$$ \mathbf{v} = \begin{bmatrix} a & b \end{bmatrix} \quad \longleftrightarrow \quad z = \begin{pmatrix} a & b \end{pmatrix}. $$

Conjugation and Negation

You can have two operations on perplex numbers that translate to discrete transformations of real 2-vectors. The first is the conjugation operation:

$$ z = \begin{pmatrix} a & b \end{pmatrix} \quad \longrightarrow \quad \operatorname{Conj}{(z)} = \begin{pmatrix} a & -b \end{pmatrix}. $$

The second is the negation operation:

$$ z = \begin{pmatrix} a & b \end{pmatrix} \quad \longrightarrow \quad \operatorname{Neg}{(z)} = \begin{pmatrix} -a & -b \end{pmatrix}. $$

Both conjugation and negation are involutions:

$$ \operatorname{Conj}{(\operatorname{Conj}{(z)})} = z; \qquad \operatorname{Neg}{(\operatorname{Neg}{(z)})} = z. $$

Subtraction

With negation and addition you can define the subtraction operation:

$$ z_{1} = \begin{pmatrix} a_{1} & b_{1} \end{pmatrix}, \qquad z_{2} = \begin{pmatrix} a_{2} & b_{2} \end{pmatrix} \quad \longrightarrow \quad \operatorname{Sub}{(z_{1}, z_{2})} = \operatorname{Add}{(z_{1}, \operatorname{Neg}{(z_{2})})} = \begin{pmatrix} a_{1} - a_{2} & b_{1} - b_{2} \end{pmatrix}. $$

Multiplication

The multiplication operation for two perplex numbers is

$$ z_{1} = \begin{pmatrix} a & b \end{pmatrix}, \qquad z_{2} = \begin{pmatrix} c & d \end{pmatrix} \quad \longrightarrow \quad \operatorname{Mul}{(z_{1}, z_{2})} = \begin{pmatrix} ac + db & da + bc \end{pmatrix}. $$

Note that

$$ \operatorname{Mul}{(z_{1}, z_{2})} = \operatorname{Mul}{(z_{2}, z_{1})}. $$

That is, the multiplication operation for perplex numbers is commutative.

Now for perplex numbers you have

$$ \operatorname{Mul}{(\operatorname{Conj}{(z_{1})}, z_{2})} = \begin{pmatrix} ac - db & da - bc \end{pmatrix}. $$

Just as for complex numbers, you can identify the real and unreal parts of this perplex number with the dot and cross products of two real 2-dimensional vectors:

$$ \operatorname{Dot}{(\mathbf{v}, \mathbf{w})} = \operatorname{Real}{(\operatorname{Mul}{(\operatorname{Conj}{(z_{1})}, z_{2})})}, \qquad \operatorname{Cross}{(\mathbf{v}, \mathbf{w})} = \operatorname{Unreal}{(\operatorname{Mul}{(\operatorname{Conj}{(z_{1})}, z_{2})})}. $$

The cross product is the same as with complex numbers,

$$ \mathbf{v} = \begin{bmatrix} v_{1} & v_{2} \end{bmatrix}, \qquad \mathbf{w} = \begin{bmatrix} w_{1} & w_{2} \end{bmatrix} \quad \longrightarrow \quad \operatorname{Cross}{(\mathbf{v}, \mathbf{w})} = v_{1} w_{2} - v_{2} w_{1}; $$

but the dot product now implies a split metric signature:

$$ \mathbf{v} = \begin{bmatrix} v_{1} & v_{2} \end{bmatrix}, \qquad \mathbf{w} = \begin{bmatrix} w_{1} & w_{2} \end{bmatrix} \quad \longrightarrow \quad \operatorname{Dot}{(\mathbf{v}, \mathbf{w})} = v_{1} w_{1} - v_{2} w_{2}. $$

Note that

$$ \operatorname{Dot}{(\mathbf{v}, \mathbf{w})} = \operatorname{Dot}{(\mathbf{w}, \mathbf{v})}, \qquad \operatorname{Cross}{(\mathbf{v}, \mathbf{w})} = -\operatorname{Cross}{(\mathbf{w}, \mathbf{v})}. $$

Quadrance

The square magnitude of a real 2-vector is now given by

$$ \mathbf{v} = \begin{bmatrix} v_{1} & v_{2} \end{bmatrix} \quad \Longrightarrow \quad \operatorname{Quad}{(\mathbf{v})} = v_{1}^{2} - v_{2}^{2}. $$

This is known as the hyperbolic quadrance. The hyperbolic quadrance can be positive, negative or zero. A perplex number also has a quadrance:

$$ z = \begin{pmatrix} a & b \end{pmatrix} \quad \Longrightarrow \quad \operatorname{Quad}{(z)} = \operatorname{Mul}{(\operatorname{Conj}{(z)}, z)} = a^{2} - b^{2}. $$

The quadrance of a perplex number can also be positive, negative or zero.

Inversion and Division

Since the quadrance of a perplex number is a non-negative real number, in general you can define the inversion operation by

$$ \operatorname{Inv}{(z)} = \operatorname{Scale}{(\operatorname{Conj}{(z)}, 1/ \operatorname{Quad}{(z)})}. $$

However, if the quadrance is zero then the inversion operation cannot be performed. The perplex numbers with zero quadrance have the form:

$$ z = \begin{pmatrix} a & a \end{pmatrix}; $$

or

$$ z = \begin{pmatrix} b & -b \end{pmatrix}. $$

A perplex number with zero quadrance is a zero-divisor. When \(a = 1/2\) or \(b = 1/2\), you have idempotents:

$$ z^{2} = z; \qquad z \neq 1. $$

Division can be defined in terms of multiplication and inversion. The division operation is

$$ \operatorname{Quo}{(z_{1}, z_{2})} = \operatorname{Mul}{(z_{1}, \operatorname{Inv}{(z_{2})})}. $$

Since the multiplication operation is commutative, the division operation is unique:

$$\operatorname{Mul}{(z_{1}, \operatorname{Inv}{(z_{2})})} = \operatorname{Mul}{(\operatorname{Inv}{(z_{2})}, z_{1})}$$

That is, there is no difference between left or right quotients.

The division operation can be define in a way that is analogous to the multiplication operation:

$$ z_{1} = \begin{pmatrix} a & b \end{pmatrix}, \qquad z_{2} = \begin{pmatrix} c & d \end{pmatrix} \quad \longrightarrow \quad \operatorname{Quo}{(z_{1}, z_{2})} = \begin{pmatrix} \dfrac{ac + db}{c^{2} + d^{2}} & \dfrac{bc - da}{c^{2} + d^{2}} \end{pmatrix}. $$

Möbius Transformation

Given a perplex number \(z\) and four other perplex numbers \(a\), \(b\), \(c\), and \(d\), the Möbius transformation of \(z\) is given by

$$ \operatorname{Mobius}{(z, a,b,c,d)} = \frac{a z + b}{c z + d}. $$

Cross-Ratio

Given four perplex numbers \(w\), \(x\), \(y\), and \(z\), the cross-ratio is given by

$$ \operatorname{CrossRatio}{(w,x,y,z)} = \frac{(w - y)(x - z)}{(x - y)(w - z)}. $$

The cross-ratio is invariant under Möbius transformations.

Differentiation

You can use two real variables \(x\) and \(t\) to construct a perplex variable \(z\):

$$ z = x + s t = \begin{pmatrix} x & t \end{pmatrix}. $$

Similarly, you can use two real differential operators \(\partial_{x}\) and \(\partial_{t}\) to construct a perplex differential operator:

$$ \partial \equiv \frac{1}{2} \left( \partial_{x} + s \partial_{t} \right). $$

This is the perplex analog of the Wirtinger derivative. Note that

$$ \partial z = 1, \qquad \partial \operatorname{Conj}{(z)} = 0, \qquad \operatorname{Conj}{(\partial)} z = 0, \qquad \operatorname{Conj}{(\partial)} \operatorname{Conj}{(z)} = 1. $$

A more useful operator is the perplex gradient operator:

$$ \nabla \equiv \partial_{x} + s \partial_{t}. $$

In terms of real 2-vectors, you have the spacetime position 2-vector:

$$ \mathbf{r} = \begin{bmatrix} x & t \end{bmatrix}. $$

and the spacetime gradient 2-vector:

$$ \nabla = \begin{bmatrix} \partial_{x} & \partial_{t} \end{bmatrix}. $$

A function 2-vector has two real functions as components:

$$ \mathbf{F}(\mathbf{r}) = \begin{bmatrix} F_{x}(x, t) & F_{t}(x, t) \end{bmatrix}. $$

Two important spacetime derivatives are the spacetime divergence:

$$ \nabla \cdot \mathbf{F} = \partial_{x} F_{x} - \partial_{t} F_{t}; $$

and the spacetime curl:

$$ \nabla \times \mathbf{F} = \partial_{x} F_{t} - \partial_{t} F_{x}. $$

Note that for real spacetime 2-vectors, you have

$$ \nabla \cdot \mathbf{r} = 1 - 1 = 0; \qquad \nabla \times \mathbf{r} = 0. $$

Writing the function 2-vector as a perplex number,

$$ F = \begin{pmatrix} F_{x}(x, t) & F_{t}(x, t) \end{pmatrix}; $$

you have

$$ 2 \operatorname{Real}{(\operatorname{Conj}{(\partial)} F)} = \operatorname{Real}{(\operatorname{Conj}{(\nabla)} F)} = \nabla \cdot \mathbf{F}, \qquad 2 \operatorname{Unreal}{(\operatorname{Conj}{(\partial)} F)} = \operatorname{Unreal}{(\operatorname{Conj}{(\nabla)} F)} = \nabla \times \mathbf{F}. $$

Assuming that the components of \(F\) are smooth, you also have

$$ 4 \partial \operatorname{Conj}{(\partial)} F = \begin{pmatrix} \partial_{x}^{2} F_{x} - \partial_{t}^{2} F_{x} & \partial_{x}^{2} F_{t} - \partial_{t}^{2} F_{t} \end{pmatrix}. $$

Instead of the 2-dimensional Laplace operator for complex numbers, you get the 2-dimensional d'Alembert operator for perplex numbers.