---
title: Multiplicative Directed Kite
url: https://www.emergentmind.com/topics/multiplicative-directed-kite
type: topic
---

# Multiplicative Directed Kite

Searching arXiv for the cited kite papers to ground the article in the literature.
Searching for "Kites and Residuated Lattices" (arXiv:1707.00267).
“Multiplicative directed kite” is an *Editor’s term* for two technically distinct uses of the word *kite* that share an explicit multiplicative component and an explicit notion of direction. In the algebraic setting, a kite is an integral residuated lattice \(K_{I_0,I_1}(G)\) built from an integral residuated lattice \(G\), nested index sets \(I_1\subseteq I_0\), and an injective map \(f:I_1\to I_0\); its layers form a “Chinese cascade kite,” and multiplication is propagated across those layers [1707.00267]. In the differential-geometric setting, a kite is a directed geometric cycle \((\sigma,\tau)\) in a pointed manifold, where \(\sigma\) is a string from the base point and \(\tau\) is a 2-simplex; this object is the domain of a 2-dimensional twisted nonabelian multiplicative integral [1007.1250]. The common vocabulary is therefore real, but the two theories address different mathematical problems.

## 1. Terminological scope

The cited literature uses *kite* in two non-equivalent senses. One is algebraic and order-theoretic; the other is geometric and integral-theoretic. The distinction is structural rather than merely notational.

| Setting | Kite | Salient feature |
|---|---|---|
| Integral residuated lattices | \(K_{I_0,I_1}(G)=\bigsqcup_{n\ge 0} G^{I_n}\) | Multiplication, residuals, lexicographic layering |
| Nonabelian multiplicative integration | A pair \((\sigma,\tau)\) in a pointed manifold | Directed/oriented surface cycle for \(MI(\alpha,\beta\mid \sigma,\tau)\) |

In the algebraic paper, the resulting algebra “has a shape of a Chinese cascade kite,” with a broad head \(G^{I_0}\) and a tapering tail. In the geometric paper, a kite is explicitly directed: \(\tau\) carries the standard orientation of \(\Delta^2\), and \(\sigma\) respects concatenation order. This suggests that any unified discussion of a “multiplicative directed kite” must keep the algebraic and geometric meanings separate, while tracking the roles played by multiplication and direction in each case [1707.00267].

## 2. Algebraic kite construction in integral residuated lattices

An algebra \((L,\wedge,\vee,\cdot,\backslash,/,e)\) of type \((2,2,2,2,2,0)\) is a residuated lattice when \((L,\wedge,\vee)\) is a lattice, \((L,\cdot,e)\) is a monoid, and
\[
x\cdot y\le z\quad\Longleftrightarrow\quad y\le x\backslash z\quad\Longleftrightarrow\quad x\le z/y
\]
for all \(x,y,z\in L\). It is integral when \(x\le e\) for every \(x\in L\). Here \(\cdot\) is multiplication, and \(\backslash\) and \(/\) are the left and right residuals.

Starting from an integral residuated lattice \(G=(G,\wedge,\vee,\cdot,\backslash,/,e)\), sets \(I_1\subseteq I_0\), and an injective map \(f:I_1\to I_0\), one defines inductively
\[
I_{n+1}=\{\,i\in I_n : f(i)\in I_n\},\qquad n=0,1,2,\dots
\]
so that
\[
I_0\supseteq I_1\supseteq I_2\supseteq\cdots.
\]
The underlying set of the kite is
\[
K_{I_0,I_1}(G):=\bigsqcup_{n\ge 0} G^{I_n}.
\]

The order is lexicographic:
\[
(x_i:i\in I_m)<(y_i:i\in I_n)
\quad\Longleftrightarrow\quad
\begin{cases}
m>n,\ \text{or}\\
m=n\ \text{and}\ x_i\le y_i\ \text{for all }i\in I_m.
\end{cases}
\]
This makes \(K\) a lattice with top element
\[
\mathbf1=(e:e\in I_0)\in G^{I_0}.
\]

Multiplication is defined layerwise. If \((x_i)\in G^{I_m}\) and \((y_i)\in G^{I_n}\), then
\[
(x_i:i\in I_m)\cdot (y_i:i\in I_n)
=
\bigl(x_i\cdot y_{f^m(i)}\bigr)_{i\in I_{m+n}},
\]
where \(f^m\) is the \(m\)-fold iterate of \(f\). The right and left residuals are defined piecewise according to the relative sizes of the layer indices, with value \(\mathbf1\) in the strict upper case and coordinatewise residuals otherwise. Theorem 3.1 states that
\[
K_{I_0,I_1}(G)
=
\bigl(\bigsqcup_{n\ge0}G^{I_n};\,\wedge,\vee,\cdot,\backslash,/,\mathbf1\bigr)
\]
is an integral residuated lattice [1707.00267].

The layered geometry is central to the name. The layers
\[
G^{I_0}\;(=\text{head}),\;G^{I_1},\;G^{I_2},\dots
\]
are stacked so that multiplication “pushes” elements downward through the levels, and once \(I_N=\varnothing\) occurs, all higher layers collapse to the singleton \(\{\mathbf1\}\). A plausible implication is that the combinatorics of the frame \((I_0,I_1,f)\) controls both the visible shape of the algebra and the depth at which nontrivial multiplication can continue to act.

## 3. Multiplicative directedness and prelinearity

In any residuated lattice, multiplication is monotone and satisfies the adjointness relation
\[
x\cdot y\le z\quad\iff\quad y\le x\backslash z\quad\iff\quad x\le z/y.
\]
For kites, the proof of Theorem 3.1 shows that this adjointness is inherited from the base algebra \(G\). Thus the multiplicative behavior of the kite is not imposed independently; it is transported from the original integral residuated lattice through the layer system determined by \(f\) [1707.00267].

A further structural transfer is given by Proposition 3.2: the kite \(K_{I_0,I_1}(G)\) satisfies prelinearity,
\[
u\backslash v\;\vee\;v\backslash u=\mathbf1
\qquad\text{and}\qquad
u/v\;\vee\;v/u=\mathbf1,
\]
if and only if the same holds in \(G\). The exposition then states: “In particular, multiplication remains ‘directed’ in the sense that any two elements become comparable up to a multiple.”

Here “directed” is not an order-theoretic synonym for total comparability. Rather, it is tied to prelinearity and residual comparison. The statement that two elements become comparable up to a multiple identifies a multiplicative substitute for direct comparability, mediated by \(\backslash\) and \(/\). This suggests that the kite construction preserves not only basic residuation but also the specific interaction between multiplication and comparability encoded by prelinearity.

## 4. Subdirect irreducibility and classification of kites

A residuated lattice is subdirectly irreducible iff it has a least nontrivial congruence, equivalently a least normal filter. For a kite \(K_{I_0,I_1}(G)\), the decisive extra datum is a graph-style connectivity relation on \(I_0\): elements \(i,j\in I_0\) are connected if for some \(m\ge 0\), either \(f^m(i)=j\) or \(f^m(j)=i\).

Theorem 5.4 gives an exact criterion. Let \(G\) be a nontrivial integral residuated lattice and let \(K=K_{I_0,I_1}(G)\) with \(I_n\neq\varnothing\). Then the following are equivalent:

1. \(G\) is subdirectly irreducible and the connection-relation on \(I_0\) is one equivalence class, i.e. \(I_0\) is connected.
2. \(K_{I_0,I_1}(G)\) is subdirectly irreducible.

The proof outline supplied in the exposition isolates the mechanism. If \(G\) has a least nontrivial normal filter \(N\), then its lift \(N^{I_0}\subseteq G^{I_0}\subseteq K\) is a least nontrivial normal filter of \(K\), provided no smaller filter cuts the kite into disconnected pieces. Conversely, if \(K\) is subdirectly irreducible, then the head \(G^{I_0}\) forces subdirect irreducibility of \(G\), while any disconnection of \(I_0\) yields two nontrivial filters with trivial intersection [1707.00267].

For finite-dimensional kites, where \(\lvert I_0\rvert<\infty\), the subdirectly irreducible cases are classified up to isomorphism as follows:

- **Head only**: \(n=0=m-1\), giving the negative cone \(\mathbb Z^-\).
- **Antilexicographic product**: \(n=m\), \(f(i)=i+1\pmod m\).
- **One tail**: \(n<m\) but \(m=n+1\), \(f(i)=i+1\) on \(\{0,\dots,n-1\}\).

For countable infinite-dimensional kites, the exposition lists three subdirectly irreducibles, again up to isomorphism:

- **Bilateral tail**: \(I_0=I_1=\mathbb Z\), \(f(i)=i+1\).
- **One-sided tail**: \(I_0=\mathbb N\), \(I_1=\mathbb N\setminus\{0\}\), \(f(i)=i-1\).
- **Two-way infinite head**: \(I_0=I_1=\mathbb Z\), \(f(i)=i+1\).

The classification shows that subdirect irreducibility is controlled by two ingredients only: the subdirectly irreducible core \(G\) and the connectivity pattern induced by \(f\). This sharply limits the possible frame shapes once irreducibility is imposed.

## 5. Variety generation and homomorphisms of frames

Let \(\mathcal K\) be the variety generated by all kites. By Birkhoff, \(\mathcal K\) is generated by subdirectly irreducible kites. Since the only infinite-dimensional subdirectly irreducible kites are the three listed in the exposition, Section 6 shows that each of them can be embedded into a suitable direct product of finite-dimensional kites by explicit maps \(\varphi_1,\varphi_2,\varphi_3\), after quotienting by a small congruence. Theorem 6.6 concludes that the variety generated by all kites coincides with the variety generated by the finite-dimensional kites. Corollary 6.7 states that \(\mathcal K\) is the varietal join of the subvarieties generated by \(n\)-dimensional kites, \(n=1,2,\dots\) [1707.00267].

This finite-dimensional generation result is structurally strong. It says that the infinite-dimensional members do not contribute new equational content beyond what is already present in finite layers. A plausible implication is that finite-dimensional frames are sufficient for describing the variety-theoretic behavior of all kites.

The same paper also formalizes morphisms between kites through maps of frames. A frame is a triple \((I_0,I_1,f)\). A map of frames
\[
t\colon (I_0,I_1,f)\longrightarrow (J_0,J_1,g)
\]
is a function \(t:I_0\to J_0\) satisfying:

1. \(t^{-1}(J_1)=I_1\),
2. \(t^{-1}(g(J_1))=f(I_1)\),
3. for each \(i\in I_1\), \(t(f(i))=g(t(i))\).

Theorem 7.6 then associates to such a frame-map the componentwise pullback
\[
K(t)\colon K_{J_0,J_1}(G)\longrightarrow K_{I_0,I_1}(G),\qquad
(x_j)_{j\in J_n}\mapsto (x_{t(i)})_{i\in I_n},
\]
and asserts that \(K(t)\) is a homomorphism of integral residuated lattices. Conversely, every homomorphism between two kites over the same \(G\) arises essentially uniquely from such a frame-map. Thus the homomorphism theory is encoded by the frame data rather than by an additional external mechanism.

## 6. Directed geometric kites and nonabelian multiplicative integration

In “Nonabelian Multiplicative Integration on Surfaces,” a kite is a geometric cycle in a pointed manifold \((X,x_0)\), not an algebra. The ambient algebraic object is a Lie crossed module \((G,H,\partial,\triangleright)\), where \(G\) and \(H\) are Lie groups, \(\partial:H\to G\) is a Lie-group homomorphism, and \(\triangleright:G\times H\to H\) is an analytic action by group automorphisms, subject to equivariance and the Pfeiffer identity. The differential data are a connection–curvature pair \((\alpha,\beta)\), with \(\alpha\in\Omega^1(X)\otimes\mathfrak g\) and \(\beta\in\Omega^2(X)\otimes\mathfrak h\), satisfying the fake-flatness equation
\[
\partial^*(\beta)=d\alpha+\tfrac12[\alpha,\alpha].
\]

A quadrangular kite in \((X,x_0)\) is a pair \((\sigma,\tau)\) in which \(\sigma\) is a string, i.e. a finite concatenation of oriented \(1\)-simplices with \(\sigma(v_0)=x_0\) and \(\sigma(v_1)=\tau(v_0)\), and \(\tau:\Delta^2\to X\) is a smooth or piecewise-linear \(2\)-simplex. The kite is directed/oriented because \(\tau\) carries the standard orientation of \(\Delta^2\), and \(\sigma\) respects concatenation order. Its boundary is the closed string
\[
\partial(\sigma,\tau)=\sigma * \partial\tau * \sigma^{-1}.
\]

The multiplicative integral \(MI(\alpha,\beta\mid \sigma,\tau)\in H\) is defined as the limit of binary-subdivision Riemann products. The basic local term is
\[
RP_0(\alpha,\beta\mid \sigma,\tau)
=
\exp_H\bigl(\operatorname{area}(Z)\cdot \Psi_{\mathfrak h}(g)(\#\beta(z))\bigr),
\]
when \(z\) is smooth for \(\beta|_Z\), with \(Z=\tau(\Delta^2)\) and \(g=MI(\alpha\mid \sigma * \text{probe})\); if \(z\) is singular or \(\dim Z<2\), then \(RP_0=1\). After subdividing the square \(I^2\) into \(4^k\) small squares and forming the corresponding tessellation of the kite, one defines
\[
RP_k(\alpha,\beta\mid \sigma,\tau)=\prod_{i=1}^{4^k} RP_0(\alpha,\beta\mid (\sigma,\tau)\circ (\sigma_i^k,\tau_i^k)).
\]
CBH estimates imply that \(RP_k\) converges in \(H\), and
\[
MI(\alpha,\beta\mid \sigma,\tau):=\lim_{k\to\infty} RP_k(\alpha,\beta\mid \sigma,\tau).
\]

The resulting theory satisfies a 2-dimensional nonabelian Stokes theorem:
\[
\partial^*\bigl(MI(\alpha,\beta\mid \sigma,\tau)\bigr)=MI(\alpha\mid \partial(\sigma,\tau))\in G,
\]
so the image under \(\partial\) is the holonomy of \(\alpha\) around the perimeter. It also satisfies the fundamental relation \(\Psi(g)=Ad_H(h)\), where \(g=MI(\alpha\mid \partial(\sigma,\tau))\) and \(h=MI(\alpha,\beta\mid \sigma,\tau)\). In dimension three, for balloons \((\sigma,\tau)\) with \(\tau:\Delta^3\to X\), the boundary multiplicative integral over the six face-kites equals the abelian multiplicative integral of the \(3\)-curvature
\[
\gamma=d\beta+\psi_{\mathfrak h}(\alpha)(\beta)\in \Omega^3(X)\otimes \mathfrak h,
\]
which lands in the inertia subalgebra \(\mathfrak h_0=\ker \operatorname{Lie}(\partial)\). If \(\gamma=0\), then \(MI_\partial=1\) [1007.1250].

This geometric theory uses “directed” in the literal sense of orientation and concatenation, whereas the algebraic kite uses “directed” through prelinearity and comparability up to a multiple. The shared term *kite* therefore indicates a formal resemblance of shape and compositional behavior, not an identity of mathematical category. The geometric construction is motivated by twisted deformation quantization, descent for nonabelian gerbes, and related questions in nonabelian gauge theory.

Source: https://www.emergentmind.com/topics/multiplicative-directed-kite