Binary Monotonic Sequences

• Binary monotonic sequences are binary strings that are either entirely non-decreasing or non-increasing, featuring a contiguous block transition between 0s and 1s.
• They are central in extremal combinatorics and serve as models for local generation protocols using communication complexes and simplicial structures.
• Their algorithmic and structural properties, including enumeration and propagation characteristics, inform studies on nonlinear recurrences and distributed computation.

A binary monotonic sequence is a string $x$ in $\{0,1\}^n$ that is either non-decreasing or non-increasing. This class — equivalently, those sequences composed of a contiguous block of 0s followed by 1s or vice versa — serves as a central object in extremal combinatorics and the study of local language generation. Binary monotonic sequences play a pivotal role in both the dynamical behaviour of nonlinear recurrences (notably, slow-growth variants of Hofstadter-type sequences) and in the structural theory of minimal communication or “generation” protocols modeled via simplicial complexes (Fox, 2016, Hoyrup, 9 Jan 2026). The study of these sequences incorporates combinatorial, algorithmic, and topological techniques to characterize their formation, extension properties, and minimal representations.

1. Definition and Core Combinatorial Structure

Given $n \geq 1$, let $I_n = \{0, 1, \ldots, n-1\}$. The set of binary monotonic sequences of length $n$ is

$$\mathrm{Mon}_n = \big\{\,0^{n-k}1^k : 0 \leq k < n \,\big\} \cup \big\{\,1^{n-k}0^k : 0 \leq k < n \,\big\}.$$

Thus, $x \in \mathrm{Mon}_n$ precisely if $x$ is non-decreasing or non-increasing as a sequence in $\{0,1\}^n$.

Key combinatorial features include:

• $|\mathrm{Mon}_n| = 2n$.
• The dihedral group $D_{2n}$ acts transitively on $\mathrm{Mon}_n$ via rotations, reflections, and bit-complementation.
• For any $i \neq j$, the pair $(x_i, x_j)$ can be prescribed arbitrarily for some $x \in \mathrm{Mon}_n$, meaning any two coordinates are independent under extension to monotonicity (Hoyrup, 9 Jan 2026).

2. Local Generation and Communication Complexes

A language $L \subseteq A^I$ is said to be locally generated if there exists a function $f: B^J \to A^I$ with $\mathrm{im}(f) = L$ and each output $i \in I$ reads only a subset $W_f(i) \subseteq J$. The communication among outputs is encoded by the communication complex $K_f$, a simplicial complex whose simplices are those $S \subseteq I$ with non-empty intersection of windows $[f]_S = \bigcap_{i \in S} W_f(i) \neq \varnothing$. A simplicial complex $K$ generates $L$ if for some $f$ with $\mathrm{im}(f) = L$, the associated $K_f \subseteq K$.

For binary monotonic sequences, $L = \mathrm{Mon}_n$, these generation complexes are deeply constrained. Minimal such $K$ are fully classified and have geometric interpretations as cyclic intervals due to the splittability property of $\mathrm{Mon}_n$ (Hoyrup, 9 Jan 2026).

3. Classification of Minimal Generating Complexes

Minimal complexes $K$ generating $\mathrm{Mon}_n$ are characterized by the structure of their maximal simplices:

• Interval Necessity: Every maximal simplex of $K$ is a cyclic interval (i.e., a subset $\{a, a+1, ..., b\}$, indices mod $n$).
• $K_2$-family: All complexes containing at least two $(n-1)$-intervals are of the form $K = \{I_n \setminus \{a\}, I_n \setminus \{b\}\}$ for distinct $a, b \in I_n$.
• $K_5$-family: There is a unique starter at $n=5$: $K_5 = \{\{0,2\}, \{1,3\}, \{2,4\}, \{3,1\}\}$; its descendants for $n \geq 5$ are formed by inserting vertices in the gaps, yielding $K^n_{i,j}$ complexes.
• $K_8$-family: For $n \geq 8$, $K_8 = \{\{0,5\},\{2,7\},\{4,1\},\{6,3\}\}$ gives rise to another infinite family via controlled vertex insertions.
• For $n \leq 7$, further sporadic minimal complexes ($K_7$, possibly $K_6$) exist, but all minimal generating structures fall within these explicit families (Hoyrup, 9 Jan 2026).

The communication constraints enforce that any minimal $K$ contains intervals of substantial length: the minimal possible maximum simplex size $\mu(n)$ satisfies $\frac{3n+1}{4} \leq \mu(n) \leq \frac{3n}{4}$ for $n \geq 8$.

4. Algorithmic and Structural Properties

Every $x \in \mathrm{Mon}_n$ is realized via a generation procedure in which locality is governed by interval windows. The canonical generation protocol uses the language itself as input and indexes outputs by maximal intervals, minimizing the communication graph. Vertex deletion in the output projects monotonic sequences of size $n+1$ to those of size $n$, preserving the communication structure. Insertion of a new position between $i-1$ and $i$ is feasible via the old input windows, provided the new bit conforms to the monotonicity constraint.

A key combinatorial lemma guarantees local propagation: for any $0 \leq i \leq j \leq k < n$, knowing $f_i = 0$ and $f_k = 0$ in some monotonic sequence ensures the existence of a compatible $f_j = 0$ extension. This propagation restricts the possible shapes of minimal generating complexes and explains the emergence of the observed $K_2$, $K_5$, $K_8$, etc., hierarchies (Hoyrup, 9 Jan 2026).

5. Binary Monotonicity in Nonlinear Recurrences

The slow three-term sequence $B(n)$, defined recursively by

$$B(n) = B\big(n - B(n-1)\big) + B\big(n - B(n-2)\big) + B\big(n - B(n-3)\big)$$

with $B(1)=1$, $B(2)=2$, $B(3)=3$, $B(4)=4$, $B(5)=5$, provides a dynamic analogue of binary monotonicity. The sequence exhibits slowness: $B(n) - B(n-1) \in \{0,1\}$ for all $n \geq 2$. This arises through an intricate induction involving repetitions and witnesses among values, and a careful accounting of when values repeat within the sequence (Fox, 2016). No similar "slow" sequences occur for $k \geq 4$-term analogues, establishing that the three-term case is maximal in this sense.

Associated to $B(n)$ is a frequency sequence $f(m)$ (values 1 or 2), and an explicit generating function: $$F(t) = \frac{t}{1-t} + \sum_{i=1}^{\infty} \frac{t^{a_i}}{1-t^{3^i}}$$ where $a_i$ form an auxiliary sequence. Asymptotic analysis yields $B(n)/n \to 2/3$.

6. Enumeration and Extremal Parameters

The enumeration of minimal generating complexes for $\mathrm{Mon}_n$ is combinatorially explicit:

• $K_2$-family: exactly $n(n-1)/2$ complexes.
• $K_5$-family: parameterized by pairs $1 < i < j < n-1$, resulting in $\binom{n-2}{2}$ minimal complexes up to symmetry.
• $K_8$-family: parameterized by quadruples $0 \leq a_0 < a_1 < a_2 < a_3 < n$ with $a_{i+1}-a_i \geq 2$ and $a_3-a_0 \leq n-2$, growth $O(n^3)$.

The parameter

$$\mu(n) = \min_{K \vdash \mathrm{Mon}_n}\ \max_{S \in K}|S|$$

satisfies the aforementioned bounds, thus minimal generating complexes inevitably have at least one interval covering roughly 75% of $I_n$ (Hoyrup, 9 Jan 2026).

7. Broader Significance and Implications

The theory of binary monotonic sequences at the interface of combinatorics, algebraic topology (through minimal simplicial complexes), and dynamical recurrence provides a paradigm for studying structure versus locality in discrete systems. These results inform the broader area of locally generated languages and the cost of distributed computation over symmetric combinatorial objects. The specific maximality results for slow sequences, together with the classification of communication complexes, illustrate sharp thresholds in the possible complexity of local generation patterns, with plausible implications for the limitations of analogues in higher-symbol or higher-arity settings (Fox, 2016, Hoyrup, 9 Jan 2026).