More on Kester function 2

 10. The Kester Spectrum

The previous sections suggest that the Kester operator should not be studied only for the particular multiplier 2.

The natural generalisation is

K_\lambda f(x)=
\begin{cases}
\lambda f(x),&x\in C_f,\\
f(x),&x\notin C_f,
\end{cases}

where f:X\to\mathbb C, C_f denotes the continuity set of f, and

\lambda\in\mathbb C.

The original Kester operator is simply the special case

K=K_2.

This generalisation reveals that the number 2 is only one point in a much larger parameter space.

The complex parameter \lambda introduces an entire Kester spectrum.

The most important division is

|\lambda|<1,\qquad |\lambda|=1,\qquad |\lambda|>1.

These three regions correspond respectively to contraction, neutral dynamics, and expansion.

But there is an additional arithmetic distinction when

|\lambda|=1.

In that case the argument of \lambda becomes decisive.

Thus the parameter plane naturally divides into several dynamical regimes.

11. The Expanding Regime

Suppose

|\lambda|>1.

For every point x,

|K_\lambda^n f(x)|
=
|\lambda|^{N_n(x)}|f(x)|.

Therefore, if

f(x)\neq0

and

N_n(x)\to\infty,

then

|K_\lambda^n f(x)|\to\infty.

The proof is immediate.

Since

|\lambda|>1,

we have

|\lambda|^{N_n(x)}\to\infty.

Hence

|K_\lambda^n f(x)|
=
|\lambda|^{N_n(x)}|f(x)|
\to\infty.

This gives the first major dynamical principle:

\boxed{
|\lambda|>1,\quad f(x)\neq0,\quad N_n(x)\to\infty
\implies
|K_\lambda^n f(x)|\to\infty.
}

The original Kester operator belongs precisely to this class.

For \lambda=2,

K^n f(x)=2^{N_n(x)}f(x).

Thus every continuity event contributes another factor of 2.

12. The Contracting Regime

Now suppose

0<|\lambda|<1.

The exact same itinerary formula gives the opposite result:

|K_\lambda^n f(x)|
=
|\lambda|^{N_n(x)}|f(x)|.

If

N_n(x)\to\infty,

then

|\lambda|^{N_n(x)}\to0.

Consequently,

\boxed{
K_\lambda^n f(x)\to0
}

for every point satisfying

f(x)\neq0

and experiencing infinitely many continuity events.

This is particularly interesting because the topology of the continuity sets has not changed in the definition of the operator.

Only the multiplier has changed.

The same itinerary that produces exponential growth for |\lambda|>1 produces exponential decay for |\lambda|<1.

Therefore the Kester dynamics possess a kind of algebraic duality:

|\lambda|>1
\quad\leftrightarrow\quad
\text{amplification},

whereas

|\lambda|<1
\quad\leftrightarrow\quad
\text{attenuation}.

13. The Neutral Circle

The most qualitatively different regime occurs when

|\lambda|=1.

Write

\lambda=e^{i\theta}.

Then

|\lambda^{N_n(x)}|=1,

and therefore

\boxed{
|K_\lambda^n f(x)|=|f(x)|
}

for every n.

The magnitude has become an invariant.

The Kester dynamics can no longer be detected by looking at absolute value.

Instead, all dynamical information is transferred into the phase.

If

f(x)=re^{i\phi},

then

K_\lambda^n f(x)
=
re^{i(\phi+N_n(x)\theta)}.

Thus the orbit lies on

S_r=\{z\in\mathbb C:|z|=r\}.

The Kester operator has therefore transformed from an amplitude-changing system into a rotation system.

14. The Phase Itinerary

For f(x)\neq0, define the phase increment

\Delta_n(x)
=
\arg\left(
\frac{K_\lambda^n f(x)}{f(x)}
\right).

Since

\frac{K_\lambda^n f(x)}{f(x)}
=
\lambda^{N_n(x)},

we obtain

\boxed{
\Delta_n(x)
=
N_n(x)\arg(\lambda)
\pmod{2\pi}.
}

This provides a direct bridge between topology and complex dynamics.

The continuity itinerary determines the number of times the phase rotation has been applied.

The multiplier determines the angle through which the phase rotates.

Consequently,

\boxed{
\text{continuity history}
\longrightarrow
\text{rotation count}
\longrightarrow
\text{phase}.
}

This is one of the most important consequences of the complex extension.

15. Rational Rotations

Suppose

\lambda=e^{2\pi i p/q},

where p/q is in lowest terms.

Then

\lambda^q=1.

Consequently,

\lambda^{N_n(x)}

depends only on

N_n(x)\pmod q.

Thus the entire complex orbit is controlled by the continuity count modulo q.

For example, take

\lambda=i.

Then

\lambda^4=1.

If f(x)=z, repeated continuity events produce

z,\quad iz,\quad -z,\quad -iz,\quad z,\ldots

provided the point experiences a continuity event at each relevant stage.

The resulting orbit has at most four distinct states.

This suggests a new notion.

Definition: Kester phase period

For a point x with f(x)\neq0, the phase period is the smallest positive integer q, when it exists, such that

\lambda^{q}=1

and the continuity itinerary generates the corresponding complete cycle.

This period is determined jointly by the multiplier and the continuity itinerary.

16. Irrational Rotations

Now let

\lambda=e^{2\pi i\alpha},

where

\alpha\notin\mathbb Q.

Then the sequence

e^{2\pi i k\alpha}

is dense in the unit circle.

Therefore, whenever the continuity itinerary generates sufficiently many distinct integer values of N_n(x), the phase orbit can become dense on

S_{|f(x)|}.

The important point is that irrationality alone is not enough.

The point must actually accumulate infinitely many distinct continuity counts.

For example, if a point is discontinuous forever after some finite stage, then

N_n(x)

eventually stops increasing.

In that case there is no infinite phase rotation, despite the irrational value of \alpha.

Thus the correct statement is conditional:

\boxed{
\alpha\notin\mathbb Q
\text{ and }
N_n(x)\to\infty
\text{ sufficiently richly}
\implies
\overline{\{K_\lambda^n f(x)\}}
=
S_{|f(x)|}.
}

The word “sufficiently” matters.

The arithmetic of \lambda and the topology of the continuity itinerary cannot be separated.

17. A New Dynamical Invariant: The Itinerary Frequency

The integer

N_n(x)

counts how many times x has been continuous.

We can refine this by considering the frequency

\rho(x)
=
\lim_{n\to\infty}
\frac{N_n(x)}{n},

when this limit exists.

We call \rho(x) the Kester continuity frequency.

It satisfies

0\leq\rho(x)\leq1.

Three extreme cases are especially simple.

If

\rho(x)=1,

the point is continuous at asymptotically almost every iteration.

If

\rho(x)=0,

continuity events occur with asymptotically negligible frequency.

If

0<\rho(x)<1,

the point experiences an intermediate density of continuity events.

This gives a new way of measuring Kester dynamics.

18. Growth Rates from Continuity Frequency

Suppose

|\lambda|>1

and \rho(x) exists.

Since

|K_\lambda^n f(x)|
=
|\lambda|^{N_n(x)}|f(x)|,

taking logarithms gives

\log |K_\lambda^n f(x)|
=
N_n(x)\log|\lambda|
+
\log|f(x)|.

Dividing by n,

\frac1n\log |K_\lambda^n f(x)|
=
\frac{N_n(x)}n\log|\lambda|
+
\frac1n\log|f(x)|.

Assuming f(x)\neq0,

\lim_{n\to\infty}
\frac1n\log |K_\lambda^n f(x)|
=
\rho(x)\log|\lambda|.

Hence

\boxed{
\gamma(x)
=
\rho(x)\log|\lambda|
}

acts like a pointwise Lyapunov exponent.

This is a particularly useful connection with dynamical systems.

The Kester operator has therefore produced its own natural notion of a Kester Lyapunov exponent.

19. The Kester Lyapunov Exponent

Define

\boxed{
\chi_f(x)
=
\limsup_{n\to\infty}
\frac1n
\log
\left|
\frac{K_\lambda^n f(x)}{f(x)}
\right|
}

for f(x)\neq0.

Using the itinerary formula,

\frac{K_\lambda^n f(x)}{f(x)}
=
\lambda^{N_n(x)}.

Therefore,

\chi_f(x)
=
\limsup_{n\to\infty}
\frac{N_n(x)}n
\log|\lambda|.

So

\boxed{
\chi_f(x)
=
\rho^+(x)\log|\lambda|
}

where

\rho^+(x)
=
\limsup_{n\to\infty}
\frac{N_n(x)}n.

This is striking because the Lyapunov exponent is controlled entirely by the continuity itinerary.

The original function values disappear from the expression.

20. The Zero Set as a Dynamical Singularity

The point

f(x)=0

is exceptional.

Regardless of \lambda,

K_\lambda^n f(x)
=
\lambda^{N_n(x)}0
=
0.

Thus

\boxed{
f(x)=0
\implies
K_\lambda^n f(x)=0
\quad\forall n.
}

The zero set is therefore pointwise invariant.

But this does not mean that continuity at such a point is invariant.

The value remains zero while the surrounding values can change dramatically.

Consequently, the zero set acts as a kind of anchor around which the topology of the continuity sets can change.

This makes zero-valued points especially important when investigating possible oscillations

x\in C_0,\quad
x\notin C_1,\quad
x\in C_2,\quad\ldots

because the value at x itself remains fixed.

21. The Continuity Set Is Not a Passive Object

One of the deepest features of the Kester construction is that C_f is not merely an external parameter.

It is generated by f itself.

The operator uses

C_f

to modify f, but modifying f changes its continuity set.

Thus

f
\longrightarrow
C_f
\longrightarrow
K_\lambda f
\longrightarrow
C_{K_\lambda f}
\longrightarrow
K_\lambda^2f
\longrightarrow\cdots

forms a feedback loop.

The function determines the topology.

The topology determines the multiplier.

The multiplier changes the function.

The changed function creates new topology.

This feedback mechanism is what makes the Kester operator a dynamical system rather than merely a piecewise-defined transformation.

22. The Kester State Space

We can therefore regard a state as containing two pieces of information:

(f,C_f).

The operator transforms this state according to

(f,C_f)
\mapsto
\left(
(1+(\lambda-1)\mathbf1_{C_f})f,
C_{K_\lambda f}
\right).

The second component is not explicitly prescribed.

It has to be calculated from the first.

Thus the induced evolution on continuity sets is nonlinear even though the pointwise transformation of values is algebraically simple.

This is an important distinction.

The operator is pointwise multiplicative but globally topological.

23. Why the Kester Operator Is Not Ordinary Scalar Multiplication

If f were simply transformed by

f\mapsto\lambda f,

then the continuity set would not change:

C_{\lambda f}=C_f

whenever \lambda\neq0.

The Kester operator is fundamentally different.

It applies multiplication only according to the continuity structure of the function.

Therefore the transformation

f\mapsto K_\lambda f

can change the continuity set.

That means the operator is capable of changing the very criterion that determines its next action.

This self-reference is the central structural feature of the theory.

24. A Generalised Kester Family

The original factor 2 can now be viewed as one member of the family

\boxed{
K_\lambda f
=
\left(
1+(\lambda-1)\mathbf1_{C_f}
\right)f.
}

Indeed,

1+(\lambda-1)\mathbf1_{C_f}
=
\begin{cases}
\lambda,&x\in C_f,\\
1,&x\notin C_f.
\end{cases}

Hence

K_2f
=
(1+\mathbf1_{C_f})f,

which is exactly the original Kester operator.

The parameter \lambda therefore provides a natural mathematical extension rather than an unrelated modification.

25. Special Values of \lambda

Several values deserve individual attention.

\lambda=1

K_1f=f.

The operator is the identity.

Every function is fixed.

\lambda=0

K_0f(x)
=
\begin{cases}
0,&x\in C_f,\\
f(x),&x\notin C_f.
\end{cases}

Every continuity point is immediately sent to zero.

\lambda=-1

K_{-1}f(x)
=
\begin{cases}
-f(x),&x\in C_f,\\
f(x),&x\notin C_f.
\end{cases}

A continuity event reverses the phase by \pi.

\lambda=i

A continuity event rotates the value by 90^\circ.

\lambda=2

The original Kester function.

\lambda=3

The function is tripled rather than doubled at continuity points:

K_3f(x)=
\begin{cases}
3f(x),&x\in C_f,\\
f(x),&x\notin C_f.
\end{cases}

This gives a direct generalisation of the original construction.

26. The a-Kester Operator

More generally, we can define

K_a f(x)=
\begin{cases}
af(x),&x\in C_f,\\
f(x),&x\notin C_f,
\end{cases}

for any scalar a.

Then

K_a^n f(x)
=
a^{N_n(x)}f(x).

The entire itinerary theory survives unchanged.

Only the algebraic effect of each continuity event changes.

Thus the fundamental object is not necessarily the number 2.

It is the pair

\boxed{(C_f,a)}.

The original Kester function corresponds to

a=2.

27. The Kester Phase Diagram

The family can therefore be represented conceptually as follows:

\begin{array}{c|c}
\text{Parameter} & \text{Typical behaviour}\\
\hline
|\lambda|>1 & \text{Expansion}\\
|\lambda|=1,\ \lambda\text{ root of unity}
& \text{Periodic phase}\\
|\lambda|=1,\ \arg\lambda/(2\pi)\notin\mathbb Q
& \text{Quasiperiodic/dense phase}\\
0<|\lambda|<1 & \text{Contraction}\\
\lambda=0 & \text{Immediate annihilation at continuity points}\\
\lambda=1 & \text{Identity}
\end{array}

This phase diagram provides a natural framework for future research.

28. A New Question: Can the Topology Detect the Parameter?

The original Kester theory starts with the multiplier and studies the topology.

But the direction can be reversed.

Suppose one observes the sequence

f,\quad K_\lambda f,\quad K_\lambda^2f,\quad\ldots

at a particular point.

Can one reconstruct \lambda from the observed orbit?

If f(x)\neq0 and x is continuous at a known stage, then

\frac{f_{n+1}(x)}{f_n(x)}=\lambda.

Thus continuity events reveal the multiplier directly.

This suggests an inverse problem:

Given the orbit of a Kester system, can one recover both the multiplier and the continuity itinerary?

This is a substantially different direction from the original continuity problem.

29. The Inverse Kester Problem

Suppose we observe

f_n(x)=K_\lambda^n f(x).

Then

\frac{f_n(x)}{f(x)}
=
\lambda^{N_n(x)}.

If f(x)\neq0, the orbit reveals powers of \lambda.

In the expanding case,

|\lambda|>1,

the magnitude may reveal

N_n(x)

directly if |\lambda| is known.

In the unit-circle case, only phase information remains.

In the rational case, different values of N_n(x) may become indistinguishable modulo the order of \lambda.

Thus the inverse problem itself has different regimes.

30. Kester Entropy

The continuity itinerary

\omega_n(x)
=
(\mathbf1_{C_0}(x),\ldots,\mathbf1_{C_{n-1}}(x))

is a binary sequence.

This suggests importing another idea from dynamical systems.

If many different itineraries occur across the domain, one can ask how rapidly the number of distinct itinerary patterns grows.

Let

P(n)

denote the number of distinct length-n itineraries realised by points of the domain.

One may then investigate

h_K
=
\limsup_{n\to\infty}
\frac{\log P(n)}n.

This could be called the Kester itinerary entropy.

It measures the combinatorial complexity of the continuity dynamics.

A system in which every point has the same continuity history would have extremely low complexity.

A system supporting many different histories could have much greater complexity.

This opens another connection between the Kester operator and symbolic dynamics.

31. Three Layers of Kester Complexity

The developing theory can now be viewed as having three distinct layers.

Layer I: Topological complexity

This concerns

C_n=C_{K_\lambda^n f}.

Questions include:

• How do the continuity sets change?
• Can they shrink?
• Can they expand?
• Can a point repeatedly enter and leave them?
• Can C_n converge to a limiting set?

Layer II: Combinatorial complexity

This concerns

\omega_n(x).

Questions include:

• Which binary itineraries are possible?
• Can arbitrary finite words occur?
• Can every infinite binary sequence occur?
• What is the itinerary entropy?

Layer III: Algebraic and geometric complexity

This concerns

\lambda^{N_n(x)}f(x).

Questions include:

• Does the magnitude grow?
• Does it decay?
• Does the phase become periodic?
• Can the orbit become dense?
• What are the fixed and periodic states?

The striking feature is that these layers are coupled.

32. A Potential Central Principle

The entire theory may ultimately be summarised by one equation:

\boxed{
K_\lambda^n f(x)
=
\lambda^{N_n(x)}f(x).
}

Everything else follows from interpreting its two factors.

The original value

f(x)

contains the initial condition.

The exponent

N_n(x)

contains the topological history.

The multiplier

\lambda

contains the local algebraic action.

Thus:

\boxed{
\text{Initial value}
\times
\text{topological history}
\times
\text{algebraic multiplier}
=
\text{Kester state}.
}

That may be the cleanest conceptual formulation of the entire theory.

33. A New Research Programme

The Kester function has now developed beyond a single unusual definition.

The natural research programme consists of several problems.

Problem 1: Itinerary realisability

Which binary sequences

\omega(x)\in\{0,1\}^{\mathbb N}

can actually arise from a Kester function?

This is perhaps the most fundamental unresolved structural question.

Problem 2: Oscillation

Can one construct an f and x such that

x\in C_0,\quad
x\notin C_1,\quad
x\in C_2,\quad
x\notin C_3,\ldots?

If such examples exist, what restrictions apply?

Problem 3: Global continuity time

Define

\tau_{\mathrm{global}}(f)
=
\inf\{n:C_n=\mathbb R\}.

Can this quantity be any element of

\mathbb N\cup\{\infty\}?

The finite staircase constructions strongly suggest that arbitrarily large finite values occur.

Problem 4: Pointwise healing time

Define

\tau_f(x)
=
\inf\{n:x\in C_n\}.

How different can the function

x\mapsto\tau_f(x)

be?

Can it be unbounded?

Can it be discontinuous?

Can it take every value in

\mathbb N_0\cup\{\infty\}?

Problem 5: Complex orbit classification

For

\lambda\in\mathbb C,

classify the orbit

\{K_\lambda^n f(x):n\ge0\}

according to \lambda and the itinerary of x.

Problem 6: Inverse dynamics

Given the orbit, determine whether one can recover

\lambda,\qquad f(x),\qquad N_n(x).

Problem 7: Statistical dynamics

Study the frequency

\rho(x)
=
\lim_{n\to\infty}\frac{N_n(x)}n

and the resulting Kester Lyapunov exponent.

34. The Broader Perspective

The Kester operator began with a deceptively simple rule:

\text{double the function where it is continuous, otherwise leave it alone.}

Yet the self-reference creates a feedback mechanism.

The function determines its continuity set.

The continuity set determines where multiplication occurs.

Multiplication changes the function.

The new function determines a new continuity set.

And the process repeats.

The complex generalisation makes the mechanism even clearer.

For real positive multipliers, continuity events alter magnitude.

For negative multipliers, they can reverse sign.

For complex unit multipliers, they rotate phase.

For complex multipliers outside the unit circle, they simultaneously rotate and expand.

For complex multipliers inside the unit circle, they rotate and contract.

Thus one elementary definition gives rise to a family of systems exhibiting:

\boxed{
\text{growth}
\;|\;
\text{decay}
\;|\;
\text{periodicity}
\;|\;
\text{quasiperiodicity}
\;|\;
\text{fixed points}
\;|\;
\text{topological feedback}.
}

The central mathematical object remains the same:

\boxed{
N_n(x)=
\sum_{j=0}^{n-1}
\mathbf1_{C_j}(x).
}

The continuity itinerary is the hidden clock of the Kester system.

Every time the clock records a continuity event, the value is multiplied by \lambda.

Everything else—the magnitude, phase, growth rate, periodicity and asymptotic behaviour—is generated from that interaction.

35. Final Conjectural Picture

The emerging picture can therefore be stated as a broad conjectural framework.

For a Kester system

K_\lambda f
=
\left(1+(\lambda-1)\mathbf1_{C_f}\right)f,

the long-term behaviour at a point x should be understood through the pair

\boxed{
(\lambda,\omega(x)).
}

The parameter \lambda determines what one continuity event does.

The itinerary \omega(x) determines when that event occurs.

The initial value f(x) determines the orbit’s starting position.

Thus the local dynamical state is encoded by

\boxed{
f(x),\quad
\lambda,\quad
\omega(x).
}

This gives a natural hierarchy:

\boxed{
\text{function}
\rightarrow
\text{continuity topology}
\rightarrow
\text{itinerary}
\rightarrow
\text{multiplicative orbit}
\rightarrow
\text{asymptotic dynamics}.
}

The original Kester function is therefore not merely the isolated rule

f\mapsto
\begin{cases}
2f,&\text{continuous},\\
f,&\text{discontinuous}.
\end{cases}

It is the first member of a broader family of self-referential topological dynamical operators.

The major question now is no longer simply:

“What does the Kester function do?”

It becomes:

Which topological dynamical systems can be generated by a function whose own continuity structure determines its future evolution?

That is the question that could take the Kester function from an interesting construction into a genuinely systematic area of mathematical investigation.