More on Kester function

Absolutely. Here is a long, paste-ready Blogger post incorporating the first 11 major points we’ve developed, including the corrected mathematics and the new local-realisability programme.

The Kester Function: A New Dynamical Theory of Continuity, Discontinuity, Repair, and Iteration

Introduction

Consider a real-valued function

f:\mathbb R\to\mathbb R.

Let

C_f=\{x\in\mathbb R:f\text{ is continuous at }x\}

denote its continuity set.

The Kester function, or more precisely the Kester operator, is defined by

\boxed{
Kf(x)=
\begin{cases}
2f(x),&f\text{ is continuous at }x,\\[4pt]
f(x),&f\text{ is discontinuous at }x.
\end{cases}}

Equivalently,

\boxed{
Kf=(1+\mathbf1_{C_f})f.
}

At first this definition looks simple. The function is doubled precisely where it is continuous and left unchanged where it is discontinuous.

But this creates a remarkable feedback mechanism.

The operator does not merely transform the numerical values of a function. It changes the function’s continuity structure, and that new continuity structure determines how the next iteration acts.

Thus we obtain a dynamical system

f,\quad Kf,\quad K^2f,\quad K^3f,\ldots

in which the topology of the function feeds back into its algebra.

This post develops the mathematical structure of this system, including discontinuity repair, finite and infinite healing, fixed points, continuity itineraries, and the open problem of local oscillation.

1. The Definition of the Kester Operator

For every f:\mathbb R\to\mathbb R, define

C_f=\{x:f\text{ is continuous at }x\}.

Then

Kf(x)=
\begin{cases}
2f(x),&x\in C_f,\\
f(x),&x\notin C_f.
\end{cases}

The important feature is that the definition is self-referential.

To calculate Kf(x), we first have to know whether f is continuous at x.

The operator therefore depends not only on the values of f, but on the topological property of f at each point.

After applying K, we obtain a new function Kf, whose continuity set can be different from C_f.

Applying K again therefore uses a different set.

This produces

C_f,\quad C_{Kf},\quad C_{K^2f},\ldots

and the resulting evolution can be studied as a dynamical system.

2. The Kester Operator Can Repair Discontinuities

One of the first surprising discoveries is that K can actually turn a discontinuous function into a continuous one.

Let

B=\{0\}\cup\left\{\frac1n:n\in\mathbb N\right\}.

Define

f(x)=
\begin{cases}
2,&x\in B,\\
1,&x\notin B.
\end{cases}

The set B is closed and has empty interior.

At every point outside B, the function is locally constant, so

C_f=\mathbb R\setminus B.

At every point of B, the function has value 2, while arbitrarily nearby points outside B have value 1. Thus every point of B is discontinuous.

Therefore

\boxed{C_f=\mathbb R\setminus B.}

Now apply K.

If x\notin B, then x\in C_f, so

Kf(x)=2f(x)=2.

If x\in B, then x\notin C_f, so

Kf(x)=f(x)=2.

Therefore

\boxed{Kf(x)=2\quad\forall x\in\mathbb R.}

\boxed{Kf\equiv2}

and hence

\boxed{C_{Kf}=\mathbb R.}

The operator has completely repaired the discontinuity set.

3. The General Complete-Repair Construction

The previous example is not an isolated curiosity.

Let B\subseteq\mathbb R be closed and nowhere dense, and define

f_B(x)=
\begin{cases}
2,&x\in B,\\
1,&x\notin B.
\end{cases}

Under the relevant condition that every point of B is approached by points outside B,

C_{f_B}=\mathbb R\setminus B.

Then

Kf_B(x)=
\begin{cases}
2(1)=2,&x\notin B,\\
2,&x\in B.
\end{cases}

Consequently,

\boxed{Kf_B\equiv2.}

Thus a whole family of discontinuous functions is transformed into a continuous constant function in one iteration.

The phenomenon can occur for complicated closed nowhere-dense sets, including familiar examples such as the Cantor set.

This demonstrates that the Kester operator can act as a discontinuity-repair mechanism.

4. The Local Mechanism of Repair

Suppose

x\notin C_f.

Then

Kf(x)=f(x).

For Kf to nevertheless become continuous at x, the surrounding values must converge to f(x).

But points approaching x can come from either the continuity set C_f or its complement.

On C_f,

Kf(t)=2f(t),

while on C_f^c,

Kf(t)=f(t).

Therefore, if both sets accumulate at x, a sufficient local criterion is

\boxed{
\lim_{\substack{t\to x\\t\in C_f}}
f(t)=\frac12f(x)
}

together with

\boxed{
\lim_{\substack{t\to x\\t\notin C_f}}
f(t)=f(x).
}

The mechanism is especially elegant.

The two different limiting values

\frac12f(x)

and

f(x)

are transformed by K into

2\left(\frac12f(x)\right)=f(x)

and

f(x).

The two branches therefore meet.

This is the fundamental algebraic mechanism behind Kester repair.

5. The Original One-Step Conjecture Is False

It was initially tempting to conjecture that the Kester operator could not repair an existing discontinuity.

That would amount to something like

C_{Kf}\subseteq C_f.

But the complete-repair example immediately disproves this.

We have

C_f=\mathbb R\setminus B

while

C_{Kf}=\mathbb R.

Therefore

\boxed{
C_{Kf}\not\subseteq C_f
}

in general.

The operator can create new discontinuities, but it can also eliminate old ones.

Thus the continuity set can move in either direction.

6. The Kester Operator as a Dynamical System

Now define

f_0=f,

and recursively

\boxed{
f_{n+1}=Kf_n.
}

Let

C_n=C_{f_n}.

Then

f_{n+1}(x)=
\begin{cases}
2f_n(x),&x\in C_n,\\
f_n(x),&x\notin C_n.
\end{cases}

The important point is that C_n itself depends on f_n.

Therefore the system has feedback:

f_n
\longrightarrow
C_n
\longrightarrow
f_{n+1}
\longrightarrow
C_{n+1}
\longrightarrow\cdots

The continuity structure controls the numerical evolution, while the numerical evolution changes the continuity structure.

This is why the operator is more naturally viewed as a dynamical system than as a single transformation.

7. The Foundational Itinerary Formula

The central bookkeeping result is the following.

Define

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

Thus N_n(x) counts the number of iterations among

0,1,\ldots,n-1

during which x was a continuity point.

Then

\boxed{
f_n(x)=2^{N_n(x)}f(x).
}

Proof

For n=1,

f_1(x)=
2^{\mathbf1_{C_0}(x)}f(x).

Assume

f_n(x)=2^{N_n(x)}f(x).

Then

f_{n+1}(x)
=
2^{\mathbf1_{C_n}(x)}f_n(x).

Substituting,

f_{n+1}(x)
=
2^{\mathbf1_{C_n}(x)}
2^{N_n(x)}f(x).

Hence

f_{n+1}(x)
=
2^{N_n(x)+\mathbf1_{C_n}(x)}f(x).

But

N_{n+1}(x)
=
N_n(x)+\mathbf1_{C_n}(x).

Therefore

f_{n+1}(x)=2^{N_{n+1}(x)}f(x).

Thus

\boxed{
f_n(x)=2^{N_n(x)}f(x)
}

for every n\ge1.

This formula is the basic algebraic invariant of Kester dynamics.

8. Continuity Itineraries

The exponent N_n(x) can be encoded by a binary string.

Define

\omega_n(x)
=
\left(
\mathbf1_{C_0}(x),
\mathbf1_{C_1}(x),
\ldots,
\mathbf1_{C_{n-1}}(x)
\right).

We call this the finite continuity itinerary of x.

For example,

\omega_3(x)=(1,0,1)

means

x\in C_0,
\qquad
x\notin C_1,
\qquad
x\in C_2.

Its Hamming weight is

N_3(x)=1+0+1=2.

Therefore

f_3(x)=4f(x).

The itinerary records the entire continuity history of a point.

This is important because two points can have the same value at one particular stage but completely different dynamical histories.

9. Arbitrarily Long Finite Kester Staircases

The next major result is that continuity repair need not happen in a single step.

For every positive integer m, we can construct a function requiring exactly m iterations to become globally continuous.

Choose distinct isolated points

x_1,\ldots,x_m

and define

f(x)=
\begin{cases}
2^j,&x=x_j,\\
1,&x\notin\{x_1,\ldots,x_m\}.
\end{cases}

Initially,

C_0=\mathbb R\setminus\{x_1,\ldots,x_m\}.

The ordinary background doubles at each iteration.

At step one, the background is

Therefore x_1, whose value is 2, becomes continuous.

At step two, the background becomes

and x_2, whose value is 4, becomes continuous.

Continuing in this way,

\boxed{
C_k=
\mathbb R\setminus
\{x_{k+1},\ldots,x_m\}.
}

Thus

\boxed{
C_0\subsetneq C_1\subsetneq\cdots\subsetneq C_m=\mathbb R.
}

Therefore:

For every m, there exists a Kester function whose discontinuities are repaired one layer at a time over exactly m iterations.

This disproves the idea that Kester repair is necessarily a one-step phenomenon.

10. Infinite Pointwise Healing

The finite staircase can be extended indefinitely.

Define

f(x)=
\begin{cases}
2^n,&x=n\in\mathbb N,\\
1,&x\notin\mathbb N.
\end{cases}

Initially,

C_0=\mathbb R\setminus\mathbb N.

The background doubles at each step.

At step k, the background has value

2^k.

Consequently the point n becomes continuous at step n.

At every finite stage k,

C_k=
\mathbb R\setminus\{k+1,k+2,k+3,\ldots\}.

Hence

\boxed{C_k\ne\mathbb R}

for every finite k.

Nevertheless, every individual point eventually becomes continuous.

Indeed, every non-integer is continuous from the beginning, while the integer n becomes continuous at stage n.

Thus

\boxed{
\bigcup_{k=0}^{\infty}C_k=\mathbb R.
}

This produces an important distinction:

\boxed{
\text{pointwise eventual continuity}
\neq
\text{global eventual continuity}.
}

Every point can eventually heal without there ever being a finite moment at which the entire function is continuous.

11. Fixed Points: Thomae’s Function

Not every Kester trajectory heals.

Consider Thomae’s function

T(x)=
\begin{cases}
\dfrac1q,&x=\dfrac pq\in\mathbb Q
\text{ in lowest terms},\\[6pt]
0,&x\notin\mathbb Q.
\end{cases}

Its continuity set is

C_T=\mathbb R\setminus\mathbb Q,

while

C_T^c=\mathbb Q.

At every irrational point,

T(x)=0.

Since irrational points are continuity points, Kester doubles the value:

KT(x)=2T(x)=0=T(x).

At rational points, T is discontinuous, so Kester leaves the value unchanged:

KT(x)=T(x).

Therefore

\boxed{KT=T.}

Consequently,

\boxed{K^nT=T\qquad\forall n\ge1.}

Thus the Kester operator possesses nonconstant fixed points.

This is especially interesting because both the continuity set and discontinuity set are dense:

\mathbb R\setminus\mathbb Q

is dense, and

\mathbb Q

is also dense.

So dense intertwining of continuity and discontinuity can produce a completely frozen Kester state.

12. The Pointwise Healing Time

The previous examples suggest a useful new quantity.

For a point x, define its Kester healing time

\boxed{
\tau_f(x)
=
\inf\{n\ge0:x\in C_n\}.
}

If x never becomes continuous, define

\tau_f(x)=\infty.

Thus

\tau_f(x)=0

means x was continuous initially.

\tau_f(x)=5,

then x first becomes continuous at iteration 5.

For a globally eventually continuous function,

\sup_x\tau_f(x)<\infty.

For the infinite staircase,

\tau_f(n)=n

for positive integers n, while

\tau_f(x)=0

for x\notin\mathbb N.

This gives us a geographical picture of the dynamics: different points can have different healing times.

13. The Possibility of Oscillation

The most difficult remaining question is whether a point can repeatedly enter and leave the continuity set.

We would like to know whether there can exist f and x such that

x\in C_0,

but

x\notin C_1,

then

x\in C_2,

and perhaps

x\notin C_3,

giving an itinerary

\boxed{(1,0,1,0,\ldots)}.

This question must be treated carefully.

It is not enough to observe that values only get multiplied by 1 or 2.

Nearby points may have different continuity histories.

A point that was discontinuous at one stage can become continuous at a later stage and consequently start doubling.

Thus its future value can “catch up” with another branch.

For this reason, simple monotonicity arguments are insufficient.

14. The Four Itinerary Classes

Suppose we seek

0\in C_0,\qquad
0\notin C_1,\qquad
0\in C_2.

Let

f(0)=a.

Since 0\in C_0,

f_1(0)=2a.

Since 0\notin C_1,

f_2(0)=2a.

For 0 to become continuous again at step two, we require

f_2(t)\to2a

as t\to0.

But nearby points can have four possible two-step itineraries:

(1,1),\quad
(1,0),\quad
(0,1),\quad
(0,0).

Their multipliers after two iterations are respectively

4,\quad2,\quad2,\quad1.

Therefore the required asymptotic behaviour is

\boxed{
\begin{array}{c|c}
\text{Itinerary}&\text{Multiplier}\\
\hline
(1,1)&4\\
(1,0)&2\\
(0,1)&2\\
(0,0)&1
\end{array}}

For continuity at 0, every branch that accumulates at 0 must ultimately produce

2a.

Thus, formally, we require

2^{N_2(t)}f(t)\to2a.

This is the exact local realisability problem.

15. The Local Realisability Problem

We can now formulate the central open problem more precisely.

Local Realisability Problem

Does there exist a function f:\mathbb R\to\mathbb R and a point x satisfying

x\in C_0,
\qquad
x\notin C_1,
\qquad
x\in C_2?

If yes, then continuity can genuinely oscillate.

If no, there must be a structural obstruction preventing the itinerary

(1,0,1).

The difficulty is that the four itinerary classes cannot simply be assigned arbitrarily.

The set C_1 is determined by the function

f_1=Kf,

and f_1 itself depends on C_0.

Therefore there is a recursive relationship

C_0
\longrightarrow f_1
\longrightarrow C_1
\longrightarrow f_2
\longrightarrow C_2.

A successful construction must satisfy all of these conditions simultaneously.

16. Why Simple Sequence Attempts Fail

A natural attempt is to choose

x_n=\frac1n\to0

and make each x_n a discontinuity while keeping 0 continuous.

For example,

f(0)=1,

f(x_n)=1+\frac1n,

and set

f(x)=1

elsewhere.

Then

f(x_n)\to1=f(0),

so 0 can indeed remain continuous even though every x_n is discontinuous.

After one Kester iteration, however, the ordinary background becomes

while the exceptional sequence remains near

Thus 0 becomes discontinuous.

At the second iteration the background becomes

while the exceptional sequence remains near 1.

The discrepancy therefore becomes even larger.

This demonstrates that a naive one-layer construction does not produce re-entry.

The surrounding points must have different continuity itineraries, not merely different initial values.

17. The Four-Layer Construction

The natural next experiment is therefore to construct four sequences

A_{11},\quad A_{10},\quad A_{01},\quad A_{00}

all accumulating at the target point 0.

We would attempt to arrange

A_{11}\subseteq C_0\cap C_1,

A_{10}\subseteq C_0\cap C_1^c,

A_{01}\subseteq C_0^c\cap C_1,

and

A_{00}\subseteq C_0^c\cap C_1^c.

Their initial values would then need to be chosen so that after two Kester iterations they all approach the same target value.

f(0)=a,

then the required asymptotic values are

\frac a2

for the (1,1) branch,

for the (1,0) and (0,1) branches,

and

for the (0,0) branch.

Thus:

\boxed{
\begin{array}{c|c}
\omega_2(t)&\text{desired asymptotic }f(t)\\
\hline
(1,1)&a/2\\
(1,0)&a\\
(0,1)&a\\
(0,0)&2a
\end{array}}

This is the first explicit algebraic blueprint for constructing a possible oscillatory Kester point.

18. Why This Is More Difficult Than It Looks

The difficulty is that we cannot simply declare

A_{11}\subseteq C_0\cap C_1

and so forth.

Continuity is a property of the whole neighbourhood.

For example, if a sequence is supposed to consist of points of C_1, we have to check that the function

f_1=Kf

is actually continuous at those points.

But f_1 itself depends on whether nearby points belong to C_0.

Thus the construction is recursive.

This makes the problem much closer to constructing a self-consistent topological dynamical system than simply defining a piecewise function.

19. A General Dynamical Formula

The itinerary formula can be rewritten as

\boxed{
K^nf(x)
=
2^{\sum_{j=0}^{n-1}\mathbf1_{C_j}(x)}f(x).
}

Therefore the Kester system separates into two components:

Numerical evolution

f(x)\longmapsto
2^{N_n(x)}f(x).

Topological evolution

C_0\longmapsto C_1\longmapsto C_2\longmapsto\cdots.

The two are coupled.

This is arguably the central conceptual feature of the Kester operator.

20. Three Proven Dynamical Regimes

At this stage we can identify at least three rigorously established behaviours.

Regime I: Immediate healing

There are discontinuous functions with

C_0\ne\mathbb R

but

C_1=\mathbb R.

The closed nowhere-dense construction provides examples.

Regime II: Finite-time healing

For every

m\in\mathbb N

there are functions with

C_0\subsetneq C_1\subsetneq\cdots\subsetneq C_m=\mathbb R.

Thus arbitrarily long finite healing times occur.

Regime III: Infinite pointwise healing

There are functions for which

C_n\ne\mathbb R

for every finite n, but

\bigcup_{n=0}^{\infty}C_n=\mathbb R.

Thus every point eventually heals without global finite-time continuity.

Regime IV: Frozen dynamics

There are nonconstant functions satisfying

Kf=f.

Thomae’s function is a particularly elegant example.

21. The Central Open Question

The next major problem is therefore:

\boxed{
\text{Which binary continuity itineraries are dynamically realisable?}
}

Given a binary sequence

\omega=(\omega_0,\omega_1,\omega_2,\ldots),

can we find a function f and point x such that

\mathbf1_{C_n}(x)=\omega_n

for every n?

For example:

(1,1,1,1,\ldots)

is certainly possible.

A continuous function gives this itinerary.

A fixed discontinuity can give

(0,0,0,0,\ldots).

Finite healing gives itineraries such as

(0,0,0,1,1,1,\ldots).

The difficult question is whether genuinely alternating sequences such as

(1,0,1,0,1,0,\ldots)

are possible.

This gives a new way of thinking about the Kester operator:

Kester dynamics may be studied through the symbolic dynamics of continuity itineraries.

22. The Kester Healing Time as a Dynamical Observable

The function

\tau_f(x)

can be regarded as a dynamical observable.

For each point, it records how long the point takes to enter the continuity set.

For the finite staircase,

\tau_f(x_j)=j.

For the infinite staircase,

\tau_f(n)=n.

For Thomae’s function,

\tau_T(x)=
\begin{cases}
0,&x\notin\mathbb Q,\\
\infty,&x\in\mathbb Q.
\end{cases}

Thus the healing-time function can itself have a rich structure.

One can ask whether

\tau_f

is bounded, unbounded, finite everywhere, infinite on a dense set, or has other topological properties.

23. A Potential Classification Programme

The Kester system now suggests a natural classification.

Class I — Continuous states

C_f=\mathbb R.

Then

Kf=2f

and consequently

\boxed{K^nf=2^nf.}

The continuity structure is permanently frozen.

Class II — Fixed discontinuous states

Kf=f.

Thomae’s function belongs here.

Class III — Finite-time healed states

There exists m<\infty such that

C_m=\mathbb R.

The finite staircase examples show that every finite m is possible.

Class IV — Pointwise but not globally healed states

Every point eventually becomes continuous, but

C_n\ne\mathbb R

for every finite n.

The integer staircase provides an example.

Class V — Persistent states

Some points never become continuous.

These include fixed points, but potentially much more complicated dynamics.

Class VI — Oscillatory states

There exists x for which the sequence

\mathbf1_{C_0}(x),
\mathbf1_{C_1}(x),
\mathbf1_{C_2}(x),\ldots

contains infinitely many transitions between 0 and 1.

Whether such states exist remains a central question.

24. A Warning About Unproved Claims

It is tempting to say that a nonzero point can only leave the continuity set once.

That has not been established.

The fact that

K^nf(x)=2^{N_n(x)}f(x)

does not by itself imply monotonicity of continuity.

A point can have one itinerary while nearby points have completely different itineraries.

For example, a nearby point may have remained discontinuous for several iterations and then become continuous, causing its value to double later.

Therefore the local neighbourhood can change its relative scaling structure over time.

Any theorem about oscillation must account for these changing itineraries.

This is why the Local Realisability Problem is so important.

25. Future Generalisation

Once the binary Kester system is understood, a natural generalisation is to replace the factor 2 with a weight function.

For example,

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

The original Kester operator corresponds to

g(x)=2.

This generalized system would introduce a richer multiplicative cocycle.

However, it is better to understand the original binary system first.

The simplest operator may already contain surprisingly complicated dynamics.

26. The Main Mathematical Picture

The Kester operator can now be viewed schematically as

\boxed{
f
\longrightarrow C_f
\longrightarrow Kf
\longrightarrow C_{Kf}
\longrightarrow K^2f
\longrightarrow\cdots
}

while at each individual point

\boxed{
f(x)
\longrightarrow
2^{N_1(x)}f(x)
\longrightarrow
2^{N_2(x)}f(x)
\longrightarrow\cdots
}

The numerical evolution is controlled by the continuity history.

The continuity history is itself generated by the numerical evolution.

That feedback is the essence of Kester dynamics.

27. What Has Been Established

The theory now contains several concrete results:

Definition of the Kester operator
Kf=(1+\mathbf1_{C_f})f.
Discontinuity repair
— K can turn a discontinuous function into a continuous one.
Complete repair families
— functions supported on closed nowhere-dense discontinuity structures can be completely repaired in one step under the stated construction.
The itinerary formula
K^nf(x)=2^{N_n(x)}f(x).
Finite Kester staircases
— arbitrarily many distinct healing stages are possible.
Infinite pointwise healing
— every point can eventually become continuous without global continuity occurring at any finite stage.
Nonconstant fixed points
— Thomae’s function satisfies
KT=T.
Pointwise healing time
\tau_f(x).
Finite continuity itineraries
\omega_n(x)\in\{0,1\}^n.
The four-class structure at two iterations
(1,1),(1,0),(0,1),(0,0).
The Local Realisability Problem
— determining which continuity itineraries can actually occur.

These results establish that the Kester operator is considerably richer than its deceptively simple definition suggests.

28. The Next Goal

The immediate mathematical objective is now very precise:

\boxed{
\text{Determine whether the itinerary }(1,0,1)
\text{ is realisable.}
}

That means finding a function f and a point x such that

x\in C_f,

x\notin C_{Kf},

and

x\in C_{K^2f}.

If such a construction exists, it establishes genuine local oscillation.

If it does not exist, then we should seek a theorem explaining why.

Either outcome would tell us something fundamental about the Kester operator.

The natural construction to investigate is a nested, multi-layer topological structure in which different sequences approaching the target point carry different continuity itineraries.

The central equation is

\boxed{
2^{N_2(t)}f(t)\longrightarrow2f(x).
}

The four possible multipliers

1,\;2,\;2,\;4

must somehow be coordinated with the local topology of the continuity sets.

That is the next frontier of the Kester-function programme.

Conclusion

The Kester function began with a simple rule:

\boxed{
Kf(x)=
\begin{cases}
2f(x),&f\text{ continuous at }x,\\
f(x),&f\text{ discontinuous at }x.
\end{cases}}

Yet this simple rule produces a feedback system in which continuity determines multiplication, multiplication changes the function, and the changed function determines a new continuity set.

We have already seen:

\text{discontinuous}
\longrightarrow
\text{continuous},

\text{discontinuous}
\longrightarrow
\text{partially healed}
\longrightarrow
\text{continuous},

\text{never globally continuous but pointwise healing},

and

\text{nonconstant fixed dynamics}.

The central algebraic identity

\boxed{
K^nf(x)=2^{N_n(x)}f(x)
}

provides the bookkeeping mechanism for all of these behaviours.

The next question is deeper:

\boxed{
\text{What continuity histories are actually possible?}
}

If arbitrary binary itineraries can be realised, Kester dynamics may possess a surprisingly rich symbolic structure.

If only certain itineraries are possible, the restrictions themselves may reveal a fundamental topological law governing the operator.

Either way, the study of the Kester function has moved beyond a single definition into the beginning of a broader theory of self-referential continuity dynamics.

Search This Blog

immigrationreformandotherpolicy

More on Kester function

Comments

Post a Comment