---
title: Truncated Riordan Groups
url: https://www.emergentmind.com/topics/truncated-riordan-groups
type: topic
---

# Truncated Riordan Groups

Searching arXiv for recent papers on Riordan groups and truncation-related constructions.
Truncated Riordan groups are finite-level versions of the Riordan group obtained by applying a natural truncation homomorphism to infinite lower-triangular Riordan matrices. For a commutative ring with identity \(\mathbb D\), the Riordan group \(\mathcal R(\mathbb D)\) consists of pairs \((g(t),f(t))\) with \(g_0\in\mathbb D^*\) and \(f_1\in\mathbb D^*\), equipped with the product
\[
(g_1(t),f_1(t))\cdot(g_2(t),f_2(t))=\bigl(g_1(t)\,g_2(f_1(t)),\,f_2(f_1(t))\bigr),
\]
and inverse
\[
(g(t),f(t))^{-1}=\left(\frac{1}{g(\bar f(t))},\,\bar f(t)\right),
\]
where \(\bar f\) is the compositional inverse of \(f\). The truncated group at level \(n\) is the image of \(\mathcal R(\mathbb D)\) under the map taking the northwest \((n+1)\times(n+1)\) block of the infinite Riordan matrix, and the full Riordan group is the inverse limit of these finite-level objects [2508.03056, 1706.01323].

## 1. Definition by finite matrix truncation

A Riordan array associated with \((g(t),f(t))\) is the infinite lower-triangular matrix \((d_{n,k})_{n,k\ge 0}\) with
\[
d_{n,k}=[t^n]\bigl(g(t)f(t)^k\bigr).
\]
For each \(n\in\mathbb N\), there is a truncation homomorphism
\[
\Pi_n:\mathcal R(\mathbb D)\longrightarrow \mathrm{GL}(n+1,\mathbb D),
\]
defined by taking the northwest \((n+1)\times(n+1)\) block. One paper writes the image as
\[
\mathcal R_n:=\Pi_n(\mathcal R),
\]
while another uses
\[
T\mathcal R_n(\mathbb D):=\Pi_n(\mathcal R(\mathbb D)).
\]
In both notations, truncated Riordan groups are finite-level lower-triangular matrix groups determined by the original pair of series [2508.03056, 2511.21639].

The transition maps between consecutive levels are given by deleting the last row and column:
\[
P_n:\mathcal R_{n+1}\to \mathcal R_n,\qquad
P_n\bigl((d_{i,j})_{0\le i,j\le n+1}\bigr)=(d_{i,j})_{0\le i,j\le n}.
\]
These satisfy
\[
\Pi_n=P_n\circ \Pi_{n+1},
\]
and \(P_n\) is surjective. Accordingly,
\[
\mathcal R(\mathbb D)\cong \varprojlim (\mathcal R_{n+1},P_n),
\]
so the infinite Riordan group is recovered as the inverse limit of its truncations [2508.03056, 1706.01323].

This matrix truncation has an equivalent power-series interpretation. Working at level \(n\) amounts to retaining only coefficients up to degree \(n\), or equivalently working modulo \(t^{n+1}\). The group law survives because multiplication and substitution of formal power series are meaningful modulo \(t^{n+1}\) [2508.03056].

## 2. Split extensions and solvability at finite level

The internal structure of truncated Riordan groups is organized by the classical Appell and Lagrange subgroups. Inside \(\mathcal R(\mathbb D)\),
\[
\mathcal A(\mathbb D)=\{(g(t),t)\mid g_0\in\mathbb D^*\}
\]
is abelian and normal, while
\[
\mathcal L(\mathbb D)=\{(1,f(t))\mid f_1\in\mathbb D^*\}
\]
is the substitution part. At each truncation level,
\[
\mathcal R_n\cong \mathcal A_n\ltimes \mathcal L_n,
\]
with \(\mathcal A_n\) and \(\mathcal L_n\) the corresponding truncated Appell and Lagrange subgroups [2508.03056].

The kernels of the transition maps control the inductive structure. For \(n=0\),
\[
\ker(P_0)\cong \mathbb D\rtimes_{\varphi}\mathbb D^*,
\qquad \varphi(a)(b)=ab.
\]
For \(n\ge 1\),
\[
\ker(P_n)\cong \mathbb D\times \mathbb D.
\]
Equivalently, for \(n\ge 1\) there are short exact sequences
\[
1\longrightarrow \mathbb D\times\mathbb D
\longrightarrow \mathcal R_{n+1}
\longrightarrow \mathcal R_n
\longrightarrow 1.
\]
The truncated Appell subgroups satisfy
\[
\mathcal A_0\cong \mathbb D^*,\qquad
\mathcal A_n\cong \mathbb D^n\times \mathbb D^*\quad (n\ge 1),
\]
and are therefore abelian. The truncated Lagrange subgroups satisfy
\[
\mathcal L_0\cong\{e\},\qquad \mathcal L_1\cong \mathbb D^*,
\]
and for \(n\ge 1\) there are short exact sequences
\[
1\longrightarrow\mathbb D
\longrightarrow \mathcal L_{n+1}
\longrightarrow \mathcal L_n
\longrightarrow 1.
\]
By induction, both \(\mathcal L_n\) and \(\mathcal R_n\) are solvable for all \(n\) [2508.03056].

This solvability sharply separates the truncated groups from the full infinite group. The same source contrasts the solvability of every \(\mathcal R_n\) with the non-solvability of the full \(\mathcal R\), using the existence of free subgroups in the substitution part in characteristic \(0\) and over finite fields [2508.03056].

## 3. Profinite structure and nilpotent truncations

When \(\mathbb D\) is finite, each \(T\mathcal R_n(\mathbb D)\) is finite, so
\[
\mathcal R(\mathbb D)\cong \varprojlim \bigl(T\mathcal R_n(\mathbb D),P_{n-1}\bigr)
\]
is a profinite group. In the special case \(\mathbb D=\mathbb F_2\), the Riordan group is a pro-\(2\) group, and the kernels of the truncation maps form a neighborhood basis of the identity [2511.21639].

A second, complementary notion of truncation is given by lower-central-series quotients. For any group \(G\),
\[
\gamma_1(G)=G,\qquad \gamma_{n+1}(G)=[\gamma_n(G),G].
\]
In \(\mathcal R(\mathbb F_2)\), the lower central series is controlled by the Appell subgroup \(\mathcal A\) and the Nottingham subgroup \(\mathcal N(\mathbb F_2)\). If
\[
\mathcal A_n
=
\{(g(t),t): g(t)=1+\alpha_{n+1}t^{n+1}+\alpha_{n+2}t^{n+2}+\cdots\},
\]
then
\[
\mathcal A\cap \gamma_{n+1}(\mathcal R)=\mathcal A_{2n-1},
\]
and for all \(n\ge 2\),
\[
\gamma_n(\mathcal R)\cong \mathcal A_{2n-3}\ltimes \gamma_n(\mathcal N).
\]
The successive lower-central quotients are
\[
\gamma_i(\mathcal R)/\gamma_{i+1}(\mathcal R)\cong
\begin{cases}
(\mathbb Z_2)^3\times \mathbb Z_4,& i=1,\\[3pt]
(\mathbb Z_2)^4,& i>1\text{ even},\\[3pt]
(\mathbb Z_2)^6,& i>1\text{ odd}.
\end{cases}
\]
These quotients supply a canonical sequence of nilpotent truncations of the full infinite group [2511.21639].

The abelianization also has a uniform description over an arbitrary commutative ring with identity:
\[
\mathcal R(\mathbb D)^{\mathrm{ab}}
\cong
\mathbb D^*\times \mathbb D\times \mathcal N(\mathbb D)^{\mathrm{ab}}.
\]
At the truncated level,
\[
T\mathcal R_n(\mathbb D)^{\mathrm{ab}}
\cong
\mathbb D^*\times \mathbb D\times T\mathcal N_n(\mathbb D)^{\mathrm{ab}}.
\]
This makes explicit how the Appell and substitution parts contribute separately to first-order approximations of the nonabelian structure [2511.21639].

## 4. Finite subgroup restrictions and non-embeddability results

A basic mechanism links embeddings in the full Riordan group to embeddings in truncations. If \(G\) is a finite group and
\[
\mu:G\to \mathcal R
\]
is a monomorphism, then there exists \(n\) such that
\[
\Pi_n\circ \mu:G\to \mathcal R_n
\]
is also a monomorphism. Consequently, the subgroup theory of the infinite Riordan group is constrained by the solvable finite-level groups \(\mathcal R_n\) [2508.03056].

This principle yields strong non-embeddability statements. For \(n\ge 4\), there is no monomorphism
\[
S_n\to \mathcal R,
\]
and no finite non-abelian simple group can be embedded into the Riordan group over any commutative ring. The proof uses the solvability of the truncations together with the descriptions of \(\ker(P_n)\): for \(n\ge 1\), \(\ker(P_n)\cong \mathbb D\times \mathbb D\), while \(\ker(P_0)\cong \mathbb D\rtimes \mathbb D^*\) is solvable. Thus any hypothetical embedding of a finite non-abelian simple group into \(\mathcal R\) would force such a group to embed into a solvable kernel, which is impossible [2508.03056].

A common misconception is that solvability of the truncations prevents interesting finite subgroup phenomena altogether. The same paper gives an explicit counterexample: the alternating group \(A_4\) embeds into a Lagrange subgroup over
\[
\mathbb D=\mathbb Z_6[X]/\langle X^2+X+1\rangle.
\]
With
\[
u=\bigl(1,\;t/(1-3t)\bigr),\qquad
w=(1,\;Xt),
\]
one has
\[
u^2=(1,t),\qquad w^3=(1,t),\qquad (uw)^3=(1,t),
\]
and the resulting subgroup is isomorphic to \(A_4\). Its image already appears in the truncated level \(\mathcal L_2(\mathbb D)\subset \mathcal R_2(\mathbb D)\). By contrast, \(A_4\) cannot embed into a substitution group \(\mathcal J\), and hence not into a Nottingham group [2508.03056].

## 5. Explicit finite \(2\)-group structure over \(\mathbb F_2\)

Over \(\mathbb F_2\), the truncated Appell groups admit a precise invariant-factor decomposition. For each \(n\ge 1\),
\[
T\mathcal A_{n+1}\cong
\mathbb Z_{2^{r_1}}\times\mathbb Z_{2^{r_2}}\times\cdots\times \mathbb Z_{2^{r_k}},
\]
where
\[
k = \left\lfloor\frac{n+1}{2}\right\rfloor,\qquad
r_j = 1+\left\lfloor \log_2\left(\frac{n+1}{2j-1}\right) \right\rfloor.
\]
The first cases are
\[
T\mathcal A_1\cong \mathbb Z_2,\quad
T\mathcal A_2\cong \mathbb Z_4,\quad
T\mathcal A_3\cong \mathbb Z_4\times\mathbb Z_2,\quad
T\mathcal A_4\cong \mathbb Z_8\times\mathbb Z_2,\quad
T\mathcal A_5\cong \mathbb Z_8\times\mathbb Z_2\times\mathbb Z_2.
\]
Moreover, the extensions
\[
0\to\mathbb F_2\to T\mathcal A_{n+1}\xrightarrow{P_n}T\mathcal A_n\to 0
\]
split if and only if \(n\) is even [2511.21639].

The finite subgroup theory of the truncated groups is correspondingly rich. For every \(n\ge 1\), there is an embedding
\[
D_{2^{n+1}}\hookrightarrow T\mathcal R_{2^n}(\mathbb F_2).
\]
The construction uses the elements
\[
r=(1+t,t)\in T\mathcal A_{2^n},\qquad
s=(1,t+t^2+\dots+t^{2^n})\in T\mathcal N_{2^n},
\]
where \(r\) has order \(2^{n+1}\), \(s^2=1\) in the truncated group, and \(rsr=s\). Thus finite dihedral groups occur naturally inside truncated Riordan groups over \(\mathbb F_2\) [2511.21639].

These examples show that truncation does not merely collapse the Riordan group to a generic lower-triangular matrix group. The Appell–Nottingham decomposition remains visible at finite level, and its arithmetic is strong enough to control both invariant factors and explicit subgroup embeddings [2511.21639].

## 6. Related generalizations and broader formulations

Several later constructions place truncated Riordan groups into a wider family of finite-level Riordan-type objects. One direction starts from almost-Riordan arrays \((a,g,f)\), which enlarge the Riordan group by freeing the first column while keeping a Riordan core from position \((1,1)\). The normal subgroup
\[
N=\{(a,1,x)\}
\]
satisfies
\[
\mathcal aR/N\cong \mathcal R,
\]
and the higher group \(R(2)\) similarly satisfies
\[
R(2)/N(2)\cong \mathcal R.
\]
The same source explicitly states that it does not use the word “truncated,” but the construction is naturally suited to matrices that differ from a Riordan array in finitely many initial columns; restricting the extra series to finite support is presented there as a natural way to obtain finite-rank perturbations of Riordan matrices [1606.05077].

A more recent extension is the Sprugnoli group, a three-series Riordan-type group designed so that “truncation in the Riordan context means” either working modulo \(x^N\) in the quotient ring \(\mathbb K[[x]]/(x^N)\) or, equivalently, working with finite lower-triangular matrices. Because its group law, inverses, and production matrices are expressed entirely through formal multiplication, substitution, and reversion, the same formulas define truncated Sprugnoli groups and, by specialization, truncated ordinary and double Riordan groups [2605.16633].

Another neighboring development replaces power series by semi-Laurent series. That work states that it “does not develop truncations directly,” but it supplies semidirect decompositions, non-isomorphism results, and an explicit Lie bracket for generalized Riordan groups involving shifted diagonals and bi-infinite matrices. A plausible implication is that finite-degree truncations of these Laurent-type constructions can be organized in parallel with the classical truncated theory [2509.04160].

Taken together, these developments show that truncated Riordan groups are not a single isolated construction but a recurrent finite-level principle: one starts from an infinite formal group law, passes to matrix blocks or quotients modulo \(x^{N+1}\), and studies the resulting solvable, nilpotent, or otherwise structured finite approximants. In the current literature, the matrix truncations \(\mathcal R_n\) and the lower-central-series quotients \(\mathcal R/\gamma_n\) are the two most developed realizations of that principle [2508.03056, 2511.21639].

Source: https://www.emergentmind.com/topics/truncated-riordan-groups