Epichristoffel Words in Episturmian Combinatorics

• Epichristoffel words are finite episturmian analogues of Christoffel words defined as the unique Lyndon word within a morphic class.
• Their construction employs episturmian morphisms and occurrence vector criteria, linking arithmetic operations to palindromic and directive properties.
• The framework generalizes binary Christoffel properties to larger alphabets, enabling novel factorization methods and the creation of epichristoffel trees.

Epichristoffel words are the finite episturmian analogue of Christoffel words. In the recent morphic formulation, an epichristoffel class is a set of words of the form $\phi(a)$, where $\phi$ is an episturmian morphism and $a$ is a letter, and an epichristoffel word is the unique Lyndon word in such a class; its conjugates are called c-epichristoffel words. In the binary case, Christoffel words are exactly the epichristoffel words on a 2-letter alphabet, so the subject extends classical Christoffel/Sturmian combinatorics into the episturmian setting over alphabets of size at least three (Krishnamoorthy et al., 21 Jul 2025, Reutenauer et al., 2024).

1. Definitions and formal setting

Christoffel words provide the binary model. Over $A=\{x,y\}$, the Christoffel word of slope $\frac{a}{b}$ is obtained by encoding the lower Christoffel path from $(0,0)$ to $(b,a)$ by $x$ for a horizontal step and $y$ for a vertical step. Equivalently, a finite word $w$ over $\phi$0 is a Christoffel word if and only if it is a balanced Lyndon word, that is, balanced and primitive and lexicographically smallest in its conjugacy class (Krishnamoorthy et al., 21 Jul 2025).

For a finite alphabet $\phi$1, the episturmian morphisms used in the epichristoffel setting are generated by the maps

$\phi$2

$\phi$3

together with the swaps

$\phi$4

and $\phi$5 for $\phi$6. The monoid generated by all $\phi$7, $\phi$8, and $\phi$9 is the monoid of episturmian morphisms, while the monoid generated only by $a$0 and $a$1 is the monoid of pure episturmian morphisms (Krishnamoorthy et al., 21 Jul 2025).

An epichristoffel class is then defined as a set of words $a$2 for which there exists an episturmian morphism $a$3 and a letter $a$4 such that $a$5. An epichristoffel word is the unique Lyndon word in such a class, and a word is c-epichristoffel if it is conjugate to an epichristoffel word. If $a$6 is epichristoffel over $a$7, the vector

$a$8

is its epichristoffel $a$9-tuple (Krishnamoorthy et al., 21 Jul 2025).

2. Episturmian environment and finite-factor placement

The natural infinite ambient class is the class of episturmian words. An infinite word $A=\{x,y\}$0 is episturmian if its factor set is closed under reversal and it has at most one right special factor of each length; it is standard if all left special factors are prefixes. Standard episturmian words admit a directive word

$A=\{x,y\}$1

and palindromic prefixes defined by

$A=\{x,y\}$2

where $A=\{x,y\}$3 is palindromic right-closure. A standard episturmian word is strict when every letter of the alphabet appears infinitely often in its directive word, and on a $A=\{x,y\}$4-letter alphabet such words have factor complexity

$A=\{x,y\}$5

For $A=\{x,y\}$6, strict episturmian words are exactly the aperiodic Sturmian words (0708.4406).

Glen’s characterization of fine words places strict episturmian and strict skew episturmian words in the same lexicographic framework. For an infinite word $A=\{x,y\}$7 with $A=\{x,y\}$8, $A=\{x,y\}$9 is fine if and only if either $\frac{a}{b}$0 is strict episturmian, or

$\frac{a}{b}$1

where $\frac{a}{b}$2 is a $\frac{a}{b}$3-strict standard episturmian word with $\frac{a}{b}$4, $\frac{a}{b}$5 is a pure epistandard morphism on $\frac{a}{b}$6, and $\frac{a}{b}$7 is a non-empty suffix of $\frac{a}{b}$8 for some $\frac{a}{b}$9. These strict skew episturmian words are non-recurrent, but all their finite factors are finite episturmian words (0708.4406).

This finite-factor layer is the immediate combinatorial habitat of epichristoffel words. The explicit conclusion drawn in the fine-word literature is that finite factors of strict episturmian words are finite episturmian words, and that Epichristoffel words live precisely in this finite episturmian category (0708.4406).

3. Morphic construction and directive data

A central existence criterion for epichristoffel words is expressed in terms of the epichristoffel $(0,0)$0-tuple. Paquin’s criterion uses an operator $(0,0)$1 on $(0,0)$2 and states that there exists an epichristoffel word with occurrence vector $(0,0)$3 if and only if some finite iteration of $(0,0)$4 applied to $(0,0)$5 yields a vector with exactly one coordinate equal to $(0,0)$6 and all other coordinates equal to $(0,0)$7. In the 3-letter example

$(0,0)$8

the corresponding morphic construction is

$(0,0)$9

so $(b,a)$0 is c-epichristoffel, and if $(b,a)$1, the epichristoffel representative is the lexicographically smallest conjugate

$(b,a)$2

(Krishnamoorthy et al., 21 Jul 2025).

The same paper makes the morphic mechanism explicit: every c-epichristoffel word with a dominant letter arises through repeated application of episturmian morphisms associated with that dominant letter. This realizes epichristoffel construction as a finite S-adic process over pure episturmian morphisms, with the tuple criterion giving the arithmetic side and the morphic composition giving the word itself (Krishnamoorthy et al., 21 Jul 2025).

The binary Christoffel theory supplies a precise prototype for this directive behavior. In the binary case, the palindromization map $(b,a)$3 satisfies

$(b,a)$4

so every proper Christoffel word is $(b,a)$5 for a finite directive word $(b,a)$6. The derivative theory then defines natural desubstitutions using morphisms

$(b,a)$7

or symmetrically

$(b,a)$8

and proves that derivatives of Christoffel words are again Christoffel words. The paper explicitly presents this as the 2-letter model for the morphic and derivative structure that epichristoffel words should carry in the episturmian setting (D'Aniello et al., 2016).

4. Factorization and epichristoffel trees

The most recent structural advance is the introduction of epichristoffel trees. The starting point is a contrast with the binary case: many properties of Christoffel words carry over to epichristoffel words, but many do not. In particular, an epichristoffel word cannot always be written as the product of two epichristoffel words; the explicit 3-letter example is

$(b,a)$9

which admits no factorization into two epichristoffel words. By contrast, any c-epichristoffel word of length $x$0 can be non-uniquely written as a product of two c-epichristoffel words (Krishnamoorthy et al., 21 Jul 2025).

The tree construction isolates a subclass where Christoffel-like factorization survives. Starting from a c-epichristoffel word

$x$1

and the factorization induced by the last morphic step,

$x$2

one passes to the Lyndon representatives of the conjugacy classes of $x$3 and $x$4, obtaining a root pair $x$5. The children are then defined exactly as in the Christoffel tree: $x$6 A key lemma proves that if $x$7 is a c-epichristoffel word whose factorization comes from the morphic construction, then

$x$8

are also c-epichristoffel. From this, every word appearing in an epichristoffel tree is shown to be epichristoffel (Krishnamoorthy et al., 21 Jul 2025).

The same construction has a tuple version. If $x$9 and $y$0 are the occurrence vectors of $y$1 and $y$2, one defines the vector mediant by

$y$3

Iterating mediant insertion yields a Stern–Brocot-type tree of $y$4-tuples associated with the epichristoffel tree. This tuple tree generates infinite families of epichristoffel $y$5-tuples from a single root. In the ternary case, it is used to prove that for every integer $y$6, except $y$7, there exists an epichristoffel word over $y$8 of length $y$9 in which each letter appears at least once (Krishnamoorthy et al., 21 Jul 2025).

The tree also separates two factorization regimes. If the root epichristoffel word cannot be factorized as a product of two epichristoffel words, then the right diagonal $w$0 of its epichristoffel tree consists entirely of epichristoffel words that cannot be so factorized. Conversely, if the root can be factorized into two epichristoffel words, then every word in its epichristoffel tree has such a factorization (Krishnamoorthy et al., 21 Jul 2025).

5. Binary Christoffel/Sturmian model

The binary Christoffel/Sturmian theory remains the reference model for epichristoffel structure. Finite binary balanced words are exactly finite Sturmian words, and a finite Sturmian word is bispecial when it is both left and right special. Central words are characterized by two coprime periods $w$1 and $w$2 with

$w$3

and strictly bispecial Sturmian words are exactly the central words. Primitive Christoffel words are exactly the words $w$4 with $w$5 whose maximal internal factor $w$6 is central (Fici, 2013).

Fici’s main theorem extends this from primitive Christoffel words to all Christoffel words: $w$7 Equivalently, bispecial Sturmian words are precisely the maximal internal factors of all Christoffel words. The internal factor $w$8 has the explicit form

$w$9

where $\phi$00 is central and $\phi$01; the case $\phi$02 gives the strictly bispecial, palindromic, primitive-Christoffel core, while $\phi$03 yields non-strictly bispecial words and non-primitive Christoffel words (Fici, 2013).

This binary correspondence also gives exact enumeration. If $\phi$04 is the number of bispecial Sturmian words of length $\phi$05, then

$\phi$06

where $\phi$07 is Euler’s totient. The same Christoffel description identifies minimal forbidden words for the finite Sturmian language as the words $\phi$08 such that $\phi$09 is a non-primitive Christoffel word (Fici, 2013).

For epichristoffel theory, this binary theorem functions as the structural template. The explicit suggestion made in the Christoffel/Sturmian literature is that a higher-alphabet “epi-bispecial” theory should again relate maximal internal factors to epichristoffel words, in the same way that bispecial Sturmian words are the maximal internal factors of Christoffel words (Fici, 2013).

6. Palindromic richness and repetition

Epichristoffel words inherit their broader palindromic context from episturmian and rich-word theory. A finite word $\phi$10 is rich when it has exactly $\phi$11 distinct palindromic factors, including $\phi$12, and a finite or infinite word is rich if and only if every complete return to every palindromic factor is itself a palindrome. Episturmian words are rich, and recurrent balanced rich infinite words are precisely the balanced episturmian words (0801.1656).

This has two consequences for epichristoffel study. First, it situates finite episturmian words inside the palindromically maximal regime. Second, it makes complete returns and palindromic closure central tools for analyzing finite episturmian objects. The same paper shows that recurrent balanced weakly rich words on at least three letters are necessarily periodic episturmian, and that balanced richness together with distinct letter frequencies drives the system toward Fraenkel-type periodicity (0801.1656).

Repetition theory imposes a complementary constraint. Every infinite rich word contains a square, and all rich square-free words are finite. If $\phi$13 denotes the length of a longest rich square-free word on an alphabet of size $\phi$14, then the known exact values for small alphabets are

$\phi$15

and the general bounds proved are

$\phi$16

Since episturmian words are rich, any infinite episturmian extension of an epichristoffel construction necessarily contains squares. This constrains the possible square-free finite factors in any episturmian or epichristoffel environment (Vesti, 2016).

7. Arithmetic and structural extensions

A recent arithmetic approach associates to a Christoffel word $\phi$17 over $\phi$18 its Burrows–Wheeler matrix $\phi$19, whose rows are the conjugates of $\phi$20 in decreasing lexicographic order. For fixed $\phi$21 and $\phi$22, these Christoffel matrices form a commutative subgroup of $\phi$23, and there is an explicit isomorphism

$\phi$24

with $\phi$25, where $\phi$26. Their determinants satisfy

$\phi$27

linking Christoffel combinatorics to the Zolotareff symbol. The same work explicitly proposes this as a blueprint for epichristoffel matrices and episturmian determinants in higher alphabets (Reutenauer et al., 2024).

A different binary refinement studies partitioned factors. For a Christoffel word $\phi$28 of length $\phi$29, a factor of length $\phi$30 partitioned according to a composition $\phi$31 is classified by the height profile of its $\phi$32 components. The theorem is that $\phi$33 is conjugate to a Christoffel word if and only if, for all $\phi$34 and $\phi$35, the multiset of $\phi$36-partitioned circular factors admits exactly $\phi$37 varieties. In the Sturmian case, the frequencies of the corresponding varieties are given by lengths of intervals cut out by the points $\phi$38 on the circle (Carey et al., 2018).

These arithmetic and factor-statistical frameworks are still binary, but they identify two directions that are directly relevant to epichristoffel words. One is matrix-theoretic and determinant-based, via BW matrices and symmetric discrete interval exchanges. The other is factor-statistical, via conjugacy, circular factors, and composition-dependent varieties. Both provide concrete higher-alphabet research programs rather than completed general theories (Reutenauer et al., 2024, Carey et al., 2018).