## Pro jection representable relations on Menger $\left(2,n\right)$-semigroups 2000 Mathematics Sub ject Classification: 20N15, 08N05. Keywords: $n$ -place function, algebra of functions, Menger algebra, $\left(2,n\right)$ -semigroup.

### November 27, 2006

Abstract
In this paper relations of non-empty intersection, inclusion end equality of domains of functions for $\left(2,n\right)$  -semigroups of partial $n$  -place functions are investigated.

1 Introduction

Investigation of partial multiplace functions by algebraic methods plays an important role in modern mathematics where we consider various operations on sets of functions which are naturally defined. The basic operation for $n$  -place functions is a superposition (composition) $O$  of $n+1$  such functions, but there are some other naturally defined operations, which are also worth considering. In this paper we consider binary Mann's compositions $\oplus 1,\dots ,\oplus n$  for partial $n$  -place functions introduced in [2, which have many important applications for the studies of binary and $n$  -ary operations. Algebras of $n$  -place functions closed with respect to these compositions were investigated, for example, in [8and [12.

2 Preliminaries and notations

Let ${A}^{n}$  be the $n$  -th Cartesian product of a set $A$  . Any partial mapping from ${A}^{n}$  into $A$  is called a partial $n$  -place function. The set of all such mappings is denoted by $\mathcal{ℱ}\left({A}^{n},A\right)$  . On $\mathcal{ℱ}\left({A}^{n},A\right)$  we define the superposition (composition) of $n$  -place functions $O:\left(f,{g}_{1},\dots ,{g}_{n}\right)↦f\left[{g}_{1}\dots {g}_{n}\right]$  and $n$  binary compositions $\oplus 1,\dots ,\oplus n$  putting
 $\begin{array}{ccc}& & f\left[{g}_{1}\dots {g}_{n}\right]\left({a}_{1},\dots ,{a}_{n}\right)=f\left({g}_{1}\left({a}_{1},\dots ,{a}_{n}\right),\dots ,{g}_{n}\left({a}_{1},\dots ,{a}_{n}\right)\right),\end{array}$ (1)
 $\begin{array}{ccc}& & \left(f\oplus ig\right)\left({a}_{1},\dots ,{a}_{n}\right)=f\left({a}_{1},\dots ,{a}_{i-1},g\left({a}_{1},\dots ,{a}_{n}\right),{a}_{i+1},\dots ,{a}_{n}\right),\end{array}$ (2)
for all $f,g,{g}_{1},\dots ,{g}_{n}\in \mathcal{ℱ}\left({A}^{n},A\right)$  and $\left({a}_{1},\dots ,{a}_{n}\right)\in {A}^{n}$  , where left and right side of (1 ) and (2 ) are defined or not defined simultaneously.
Since, as it is not difficult to verify, each composition $\oplus i$  is an associative operation, algebras of the form $\left(\Phi ;\oplus 1,\dots ,\oplus n\right)$  and $\left(\Phi ;O,\oplus 1,\dots ,\oplus n\right)$  , where $\Phi \subset \mathcal{ℱ}\left({A}^{n},A\right)$  , are called respectively $\left(2,n\right)$  -semigroups and Menger $\left(2,n\right)$  -semigroups of $n$  -place functions.
According to the general convention used in the theory of $n$  -ary systems, the sequence ${x}_{i},{x}_{i+1},\dots ,{x}_{j}$  , where $i⩽j$  , can be written as ${x}_{i}^{j}$  (for $i>j$  it is the empty symbol). In this convention (1 ) and (2 ) can be written as
 $\begin{array}{ccc}& & f\left[{g}_{1}^{n}\right]\left({a}_{1}^{n}\right)=f\left({g}_{1}\left({a}_{1}^{n}\right),\dots ,{g}_{n}\left({a}_{1}^{n}\right)\right),\end{array}$
 $\begin{array}{ccc}& & \left(f\oplus ig\right)\left({a}_{1}^{n}\right)=f\left({a}_{1}^{i-1},g\left({a}_{1}^{n}\right),{a}_{i+1}^{n}\right).\end{array}$
An algebra $\left(G;o\right)$  with one $\left(n+1\right)$  -ary operation $o$  satisfying the identity $o\left(o\left({x}_{0}^{n}\right),{y}_{1}^{n}\right)=o\left({x}_{0},o\left({x}_{1},{y}_{1}^{n}\right),\dots ,o\left({x}_{n},{y}_{1}^{n}\right)\right)$  is called a Menger algebra of rank $n$  (cf. [1, [7). Such operation is called superassociative and by many authors is written as $o\left({x}_{0}^{n}\right)={x}_{0}\left[{x}_{1}^{n}\right]$  . In this convention the above identity has the form
 $\begin{array}{c}{x}_{0}\left[{x}_{1}^{n}\right]\left[{y}_{1}^{n}\right]={x}_{0}\left[{x}_{1}\left[{y}_{1}^{n}\right]\dots {x}_{n}\left[{y}_{1}^{n}\right]\right].\end{array}$ (3)
It is clear that a Menger algebra of rank $1$  is an arbitrary semigroup.
Let $\left\{\oplus 1,\dots ,\oplus n\right\}$  be a collection of associative binary operations defined on $G$  . According to [8and [12, an algebra $\left(G;\oplus 1,\dots ,\oplus n\right)$  is called a $\left(2,n\right)$  -semigroup. By a Menger $\left(2,n\right)$  -semigroup we mean an algebra $\left(G;o,\oplus 1,\dots ,\oplus n\right)$  , where $\left(G;o\right)$  is a Menger algebra of rank $n$  and $\left(G;\oplus 1,\dots ,\oplus n\right)$  is a $\left(2,n\right)$  -semigroup. Any homomorphism of a (Menger) $\left(2,n\right)$  -semigroup onto some (Menger) $\left(2,n\right)$  -semigroup of $n$  -place functions is called a representation by $n$  -place functions. A representation is faithful if it is an isomorphism.
The symbol ${\mu }_{i}\left(\oplus {i}_{s}{i}_{1}{x}_{1}^{s}\right)$  , where ${x}_{1},\dots ,{x}_{s}\in G$  and $\oplus {i}_{1},\dots ,\oplus {i}_{s}$  are binary operations defined on $G$  , denotes an element ${x}_{{i}_{k}}\oplus {i}_{s}{i}_{k+1}{x}_{k+1}^{s}$  if $i={i}_{k}$  and $i\ne {i}_{p}$  for all $p  . If $i\ne {i}_{p}$  for all ${i}_{p}\in \left\{{i}_{1},\dots ,{i}_{s}\right\}$  this symbol is empty. For example, ${\mu }_{1}\left(\oplus 2x\oplus 1y\oplus 3z\right)=y\oplus 3z$  , ${\mu }_{2}\left(\oplus 2x\oplus 1y\oplus 3z\right)=x\oplus 1y\oplus 3z$  , ${\mu }_{3}\left(\oplus 2x\oplus 1y\oplus 3z\right)=z$  . The symbol ${\mu }_{4}\left(\oplus 2x\oplus 1y\oplus 3z\right)$  is empty.
In [8it is proved that a $\left(2,n\right)$  -semigroup $\left(G;\oplus 1,\dots ,\oplus n\right)$  has a faithful representation by $n$  -place functions if and only if it satisfies the implication
 $\begin{array}{c}{\bigwedge }_{i=1}^{n}\left({\mu }_{i}\left(\oplus {i}_{s}{i}_{1}{x}_{1}^{s}\right)={\mu }_{i}\left(\oplus {j}_{k}{j}_{1}{y}_{1}^{k}\right)\right)⟶g\oplus {i}_{s}{i}_{1}{x}_{1}^{s}=g\oplus {j}_{k}{j}_{1}{y}_{1}^{k}.\end{array}$ (4)
For Menger $\left(2,n\right)$  -semigroups the following identities must be satisfied additionally
 $\begin{array}{ccc}& & \left(x\oplus iy\right)\left[{z}_{1}^{n}\right]=x\left[{z}_{1}^{i-1}y\left[{z}_{1}^{n}\right]{z}_{i+1}^{n}\right],\end{array}$ (5)
 $\begin{array}{ccc}& & x\left[{y}_{1}^{n}\right]\oplus iz=x\left[{y}_{1}\oplus iz\dots {y}_{n}\oplus iz\right],\end{array}$ (6)
 $\begin{array}{ccc}& & x\oplus {i}_{s}{i}_{1}{y}_{1}^{s}=x\left[{\mu }_{1}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)\dots {\mu }_{n}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)\right],\end{array}$ (7)
where $\left\{{i}_{1},\dots ,{i}_{s}\right\}=\left\{1,\dots ,n\right\}$  and $i=1,\dots ,n$  . In the sequel, any (Menger) $\left(2,n\right)$  -semigroup satisfying the condition (4 ) (respectively, (4 ), (5 ), (6 ) and (7 )) will be called representable.
Let $\Phi$  be some set of $n$  -place functions, i.e. $\Phi \subset \mathcal{ℱ}\left({A}^{n},A\right)$  . Consider the following three binary relations on $\Phi$  :
 $\begin{array}{ccc}& & {\chi }_{\Phi }=\left\{\left(f,g\right)\in \Phi ×\Phi |\text{pr}1f\subset \text{pr}1g\right\},\end{array}$
 $\begin{array}{ccc}& & {\gamma }_{\Phi }=\left\{\left(f,g\right)\in \Phi ×\Phi |\text{pr}1f\cap \text{pr}1g\ne \mathbb{\varnothing }\right\},\end{array}$
 $\begin{array}{ccc}& & {\pi }_{\Phi }=\left\{\left(f,g\right)\in \Phi ×\Phi |\text{pr}1f=\text{pr}1g\right\},\end{array}$
where $\text{pr}1f$  is the domain of $f$  , called respectively: inclusion of domains, co-definability and equality of domains.
Abstract characterizations of such relations for semigroups of transformations were studied in [4, [5, [6and for Menger algebras of $n$  -place functions in [9, [10, [11. In this paper these relations will be characterized in $\left(2,n\right)$  -semigroups and Menger $\left(2,n\right)$  -semigroups of $n$  -place functions.
Consider a representable (Menger) $\left(2,n\right)$  -semigroup $\left(G;\oplus 1,\dots ,\oplus n\right)$  (respectively, $\left(G;o,\oplus 1,\dots ,\oplus n\right)$  ) and its representation $P$  by $n$  -place functions. On the set $G$  we define the following three binary relations:
 $\begin{array}{ccc}& & {\chi }_{P}=\left\{\left({g}_{1},{g}_{2}\right)|\text{pr}1P\left({g}_{1}\right)\subset \text{pr}1P\left({g}_{2}\right)\right\},\end{array}$
 $\begin{array}{ccc}& & {\gamma }_{P}=\left\{\left({g}_{1},{g}_{2}\right)|\text{pr}1P\left({g}_{1}\right)\cap \text{pr}1P\left({g}_{2}\right)\ne \mathbb{\varnothing }\right\},\end{array}$
 $\begin{array}{ccc}& & {\pi }_{P}=\left\{\left({g}_{1},{g}_{2}\right)|\text{pr}1P\left({g}_{1}\right)=\text{pr}1P\left({g}_{2}\right)\right\}.\end{array}$
It is not difficult to see that ${\chi }_{P}$  is a quasi-order and ${\pi }_{P}$  is an equivalence such that ${\pi }_{P}={\chi }_{P}\cap {\chi }_{P}^{-1}$  , where ${\chi }_{P}^{-1}=\left\{\left(b,a\right)|\left(a,b\right)\in {\chi }_{P}\right\}$  .
Let $\left({P}_{i}{\right)}_{i\in I}$  be a family of representations of a representable $\left(2,n\right)$  -semigroup $\left(G;\oplus 1,\dots ,\oplus n\right)$  (respectively, representable Menger $\left(2,n\right)$  -semigroup $\left(G;o,\oplus 1,\dots ,\oplus n\right)$  ) by $n$  -place functions defined on sets $\left({A}_{i}{\right)}_{i\in I}$  respectively, where the sets ${A}_{i}$  are pairwise disjoint. The sum of $\left({P}_{i}{\right)}_{i\in I}$  is the mapping $P:g↦P\left(g\right)$  , denoted by ${\sum }_{i\in I}{P}_{i}$  , where $P\left(g\right)$  is an $n$  -place function on $A={\cup }_{i\in I}{A}_{i}$  such that $P\left(g\right)={\cup }_{i\in I}{P}_{i}\left(g\right)$  for every $g\in G$  . The sum of a family of representations by $n$  -place functions is also a representation by $n$  -place functions and
 $\begin{array}{c}{\chi }_{P}={\bigcap }_{i\in I}{\chi }_{{P}_{i}},{\gamma }_{P}={\bigcup }_{i\in I}{\gamma }_{{P}_{i}},{\pi }_{P}={\bigcap }_{i\in I}{\pi }_{{P}_{i}}.\end{array}$ (8)
Let $0$  be a zero of a $\left(2,n\right)$  -semigroup $\left(G;\oplus 1,\dots ,\oplus n\right)$  (respectively, Menger $\left(2,n\right)$  -semigroup $\left(G;o,\oplus 1,\dots ,\oplus n\right)$  ), i.e. $0\oplus ig=g\oplus i0=0$  (respectively, $0\oplus ig=g\oplus i0=0$  and $0\left[{g}_{1}^{n}\right]=g\left[{g}_{1}^{i-1}0{g}_{i+1}^{n}\right]=0$  ) for all $i=1,\dots ,n$  and $g,{g}_{1},\dots ,{g}_{n}\in G$  . We say that a binary relation $\rho \subset G×G$  is $0$  -reflexive, if $\left(g,g\right)\in \rho$  for all $g\in G\\left\{0\right\}$  . A symmetric relation $\rho$  which is reflexive if $0\in \text{pr}1\rho$  , and $0$  -reflexive if $0\notin \text{pr}1\rho$  , is called a $0$  -quasi-equivalence.
A binary relation $\Delta$  on a Menger $\left(2,n\right)$  -semigroup $\left(G;o,\oplus 1,\dots ,\oplus n\right)$  is called:
• $l$  -regular, if  $\begin{array}{ccc}& & x\Delta y⟶x\left[{z}_{1}^{n}\right]\Delta y\left[{z}_{1}^{n}\right],\end{array}$ (9)
 $\begin{array}{ccc}& & x\Delta y⟶x\oplus iz\Delta y\oplus iz\end{array}$ (10)
for all $i=1,\dots ,n$  and $x,y,z,{z}_{1},\dots ,{z}_{n}\in G$  ,
• $l$  -cancellative, if  $\begin{array}{ccc}& & x\left[{z}_{1}^{n}\right]\Delta y\left[{z}_{1}^{n}\right]⟶x\Delta y,\end{array}$ (11)
 $\begin{array}{ccc}& & x\oplus iz\Delta y\oplus iz⟶x\Delta y\end{array}$ (12)
for all $i=1,\dots ,n$  and $x,y,z,{z}_{1},\dots ,{z}_{n}\in G$  ,
• $v$  -negative, if  $\begin{array}{ccc}& & x\left[{y}_{1}^{n}\right]\Delta {y}_{i},i=1,\dots ,n,\end{array}$ (13)
 $\begin{array}{ccc}& & x\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\Delta {\mu }_{j}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)\end{array}$ (14)
for all $x,{y}_{1},\dots ,{y}_{k}\in G$  , $k=max\left\{n,s\right\}$  and $j\in \left\{{i}_{1},\dots ,{i}_{s}\right\}$  .
In the case of $\left(2,n\right)$  -semigroups these relations are defined only by (10 ), (12 ) and (14 ), respectively.

3 Pro jection representable relations on Menger $\left(2,n\right)$  -semigroups

Let $\mathcal{G}=\left(G;o,\oplus 1,\dots ,\oplus n\right)$  be a representable Menger $\left(2,n\right)$  -semigroup, $\chi$  , $\gamma$  , $\pi$  – binary relations on $G$  . We say that the triplet $\left(\chi ,\gamma ,\pi \right)$  is (faithful) projection representable for $\mathcal{G}$  , if there exists such (faithful) representation $P$  of $\mathcal{G}$  by $n$  -place functions for which $\chi ={\chi }_{P}$  , $\gamma ={\gamma }_{P}$  and $\pi ={\pi }_{P}$  .
Analogously we define projection representable pairs and separate relations.
In the sequel, instead of $\left({g}_{1},{g}_{2}\right)\in \chi$  , $\left({g}_{1},{g}_{2}\right)\in \gamma$  and $\left({g}_{1},{g}_{2}\right)\in \pi$  we will write ${g}_{1}⊏{g}_{2}$  , ${g}_{1}\top {g}_{2}$  and ${g}_{1}\equiv {g}_{2}$  , respectively.
Theorem 1. A triplet $\left(\chi ,\gamma ,\pi \right)$  of binary relations on $G$  is projection representable for a representable Menger $\left(2,n\right)$  -semigroup $\mathcal{G}$  if and only if the following conditions are satisfied:
$\left(a\right)$  $\chi$  is an $l$  -regular and $v$  -negative quasi-order, $\left(b\right)$  $\gamma$  is an $l$  -cancellative $0$  -quasi-equivalence, $\left(c\right)$  $\pi =\chi \cap {\chi }^{-1}$  and
 $\begin{array}{c}{h}_{1}\top {h}_{2}\wedge {h}_{1}⊏{g}_{1}\wedge {h}_{2}⊏{g}_{2}⟶{g}_{1}\top {g}_{2}\end{array}$ (15)
for all ${h}_{1},{h}_{2},{g}_{1},{g}_{2}\in G$  .
• Proof. Necessity. Let $\left(\Phi ;O,\oplus 1,\dots ,\oplus n\right)$  be a Menger $\left(2,n\right)$  -semigroup of $n$  -place functions determined on the set $A$  . Let us show that the triplet $\left({\chi }_{\Phi },{\gamma }_{\Phi },{\pi }_{\Phi }\right)$  satisfies all the conditions of the theorem.
At first we prove the condition $\left(a\right)$  . The relation ${\chi }_{\Phi }$  is obviously a quasi-order. Let $f,g,{h}_{1},\dots ,{h}_{n}\in \Phi$  and $\left(f,g\right)\in {\chi }_{\Phi }$  , i.e. $\text{pr}1f\subset \text{pr}1g$  .
Suppose that $\overline{a}\in \text{pr}1f\left[{h}_{1}^{n}\right]$  for some $\overline{a}\in {A}^{n}$  . Then $\left\{f\left[{h}_{1}^{n}\right]\left(\overline{a}\right)\right\}\ne \mathbb{\varnothing }$  , i.e.
$\left\{f\left({h}_{1}\left(\overline{a}\right),\dots {h}_{n}\left(\overline{a}\right)\right)\right\}\ne \mathbb{\varnothing }$  . Thus $\left({h}_{1}\left(\overline{a}\right),\dots {h}_{n}\left(\overline{a}\right)\right)\in \text{pr}1f$  and, in the consequence, $\left({h}_{1}\left(\overline{a}\right),\dots {h}_{n}\left(\overline{a}\right)\right)\in \text{pr}1g$  . Therefore $\left\{g\left({h}_{1}\left(\overline{a}\right),\dots {h}_{n}\left(\overline{a}\right)\right)\right\}\ne \mathbb{\varnothing }$  , whence $\left\{g\left[{h}_{1}^{n}\right]\left(\overline{a}\right)\right\}\ne \mathbb{\varnothing }$  , i.e. $\overline{a}\in \text{pr}1g\left[{h}_{1}^{n}\right]$  . So, $\text{pr}1f\left[{h}_{1}^{n}\right]\subset \text{pr}1g\left[{h}_{1}^{n}\right]$  , which implies $\left(f\left[{h}_{1}^{n}\right],g\left[{h}_{1}^{n}\right]\right)\in {\chi }_{\Phi }$  . Similarly we can prove that for all $f,g,h\in \Phi$  and $i=1,\dots ,n$  , from $\left(f,g\right)\in {\chi }_{\Phi }$  it follows $\left(f\oplus ih,g\oplus ih\right)\in {\chi }_{\Phi }$  . This means that the relation ${\chi }_{\Phi }$  is $l$  -regular. The proof of the $v$  -negativity is analogous.
To prove $\left(b\right)$  let $\Theta$  be a zero of a Menger $\left(2,n\right)$  -semigroup $\left(\Phi ;O,\oplus 1,\dots ,\oplus n\right)$  . If $\Theta \ne \mathbb{\varnothing }$  , then $\text{pr}1\Theta \ne \mathbb{\varnothing }$  , whence $\left(\Theta ,\Theta \right)\in {\gamma }_{\Phi }$  .
Thus $\Theta \in \text{pr}1{\gamma }_{\Phi }$  . So, in this case ${\gamma }_{\Phi }$  is reflexive. For $\Theta =\mathbb{\varnothing }$  we have $\text{pr}1\Theta =\mathbb{\varnothing }$  . Therefore $\Theta \notin \text{pr}1{\gamma }_{\Phi }$  , i.e. $\left(f,f\right)\in {\gamma }_{\Phi }$  for every $f\ne \Theta$  . Hence ${\gamma }_{\Phi }$  is $\Theta$  -reflexive. Since ${\gamma }_{\Phi }$  is symmetric, the above means that ${\gamma }_{\Phi }$  is a $\Theta$  -quasi-equivalence.
Suppose now that $\left(f\left[{h}_{1}^{n}\right],g\left[{h}_{1}^{n}\right]\right)\in {\gamma }_{\Phi }$  for some $f,g\in \Phi$  , ${h}_{1}^{n}\in {\Phi }^{n}$  . Then $\text{pr}1f\left[{h}_{1}^{n}\right]\cap \text{pr}1g\left[{h}_{1}^{n}\right]\ne \mathbb{\varnothing }$  , i.e. there exists $\overline{a}\in {A}^{n}$  such that $\overline{a}\in \text{pr}1f\left[{h}_{1}^{n}\right]$  and $\overline{a}\in \text{pr}1g\left[{h}_{1}^{n}\right]$  . Therefore $\left\{f\left[{h}_{1}^{n}\right]\left(\overline{a}\right)\right\}\ne \mathbb{\varnothing }$  and $\left\{g\left[{h}_{1}^{n}\right]\left(\overline{a}\right)\right\}\ne \mathbb{\varnothing }$  . Thus $\left\{f\left({h}_{1}\left(\overline{a}\right),\dots {h}_{n}\left(\overline{a}\right)\right)\right\}\ne \mathbb{\varnothing }$  and $\left\{g\left({h}_{1}\left(\overline{a}\right),\dots {h}_{n}\left(\overline{a}\right)\right)\right\}\ne \mathbb{\varnothing }$  , which shows that $\left({h}_{1}\left(\overline{a}\right),\dots ,{h}_{n}\left(\overline{a}\right)\right)\in \text{pr}1f\cap \text{pr}1g$  . So, $\left(f,g\right)\in {\gamma }_{\Phi }$  . Analogously, for $f,g,h\in \Phi$  , $i=1,\dots ,n$  , from $\left(f\oplus ih,g\oplus ih\right)\in {\gamma }_{\Phi }$  it follows $\left(f,g\right)\in {\gamma }_{\Phi }$  . So, ${\gamma }_{\Phi }$  is $l$  -cancellative.
Since in $\left(c\right)$  the first condition is obvious, we prove $\left(15\right)$  only. For this let $\left({h}_{1},{h}_{2}\right)\in {\gamma }_{\Phi }$  , $\left({h}_{1},{g}_{1}\right)\in {\chi }_{\Phi }$  and $\left({h}_{2},{g}_{2}\right)\in {\chi }_{\Phi }$  for some ${h}_{1},{h}_{2},{g}_{1},{g}_{2}\in \Phi$  .
Then $\text{pr}1{h}_{1}\cap \text{pr}1{h}_{2}\ne \mathbb{\varnothing }$  , $\text{pr}1{h}_{1}\subset \text{pr}1{g}_{1}$  and $\text{pr}1{h}_{2}\subset \text{pr}1{g}_{2}$  , whence $\mathbb{\varnothing }\ne \text{pr}1{h}_{1}\cap \text{pr}1{h}_{2}\subset \text{pr}1{g}_{1}\cap \text{pr}1{g}_{2}$  . Thus $\text{pr}1{g}_{1}\cap \text{pr}1{g}_{2}\ne \mathbb{\varnothing }$  , i.e. $\left({g}_{1},{g}_{2}\right)\in {\gamma }_{\Phi }$  , which proves (15 ) and completes the proof of the necessity of the conditions formulated in the theorem.
To prove the sufficiency of these conditions we must introduce some additional constructions. Consider the triplet $\left(\chi ,\gamma ,\pi \right)$  of binary relations on a representable Menger $\left(2,n\right)$  -semigroup $\mathcal{G}=\left(G;o,\oplus 1,\dots ,\oplus n\right)$  satisfying all the conditions of the theorem. Let ${e}_{1},\dots ,{e}_{n}$  be pairwise different elements not belonging to $G$  . For all ${x}_{1},\dots ,{x}_{s}\in G$  , $i=1,\dots ,n$  , and operations $\oplus {i}_{1},\dots ,\oplus {i}_{s}$  defined on $G$  by ${\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{x}_{1}^{s}\right)$  we denote an element of ${G}^{*}=G\cup \left\{{e}_{1},\dots ,{e}_{n}\right\}$  such that ${\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{x}_{1}^{s}\right)=\left\{\begin{array}{cc}{\mu }_{i}\left(\oplus {i}_{s}{i}_{1}{x}_{1}^{s}\right),& \text{if}i\in \left\{{i}_{1},\dots ,{i}_{s}\right\},\\ {e}_{i},& \text{if}i\notin \left\{{i}_{1},\dots ,{i}_{s}\right\}.\end{array}$  Consider the set ${\mathfrak{A}}^{*}={G}^{n}\cup {\mathfrak{A}}_{0}\cup \left\{\left({e}_{1},\dots ,{e}_{n}\right)\right\}$  , where ${\mathfrak{A}}_{0}$  is the collection of all $n$  -tuples $\left({x}_{1},\dots ,{x}_{n}\right)\in \left({G}^{*}{\right)}^{n}$  for which there exists ${y}_{1},\dots ,{y}_{s}\in G$  and ${i}_{1},\dots ,{i}_{n}\in \left\{1,\dots ,n\right\}$  such that ${x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)$  . Let $\left({h}_{1},{h}_{2}\right)\in {G}^{2}$  be fixed. For each $g\in G$  we define a partial $n$  -place function ${P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right):{\mathfrak{A}}^{*}\to G$  such that ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)⟷\left\{\begin{array}{cc}{h}_{1}⊏g\left[{x}_{1}^{n}\right]\vee {h}_{2}⊏g\left[{x}_{1}^{n}\right]& \text{if}{x}_{1}^{n}\in {G}^{n},\\ {h}_{1}⊏g\vee {h}_{2}⊏g& \text{if}{x}_{1}^{n}={e}_{1}^{n},\\ {h}_{1}⊏g\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\vee {h}_{2}⊏g\oplus {i}_{s}{i}_{1}{y}_{1}^{s}& \text{if}{x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right),\\ & i=1,\dots ,n,\text{for}\\ & \text{some}{y}_{1}^{s}\in {G}^{s}\text{and}\\ & {i}_{1}\dots ,{i}_{s}\in \left\{1,\dots ,n\right\}.\end{array}$  For ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)$  we put
 $\begin{array}{c}{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)\left({x}_{1}^{n}\right)=\left\{\begin{array}{cc}g\left[{x}_{1}^{n}\right]& \text{if}{x}_{1}^{n}\in {G}^{n},\\ g& \text{if}{x}_{1}^{n}={e}_{1}^{n},\\ g\oplus {i}_{s}{i}_{1}{y}_{1}^{s}& \text{if}{x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right),\\ & i=1,\dots ,n,\text{for}\\ & \text{some}{y}_{1}^{s}\in {G}^{s}\text{and}\\ & {i}_{1}\dots ,{i}_{s}\in \left\{1,\dots ,n\right\}.\end{array}\end{array}$ (16)
Let us show that ${P}_{\left({h}_{1},{h}_{2}\right)}$  is a representation of $\mathcal{G}$  by $n$  -place functions.
Proposition 1. The function ${P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)$  is single-valued.
• Proof. Let ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)$  , where $g,{h}_{1},{h}_{2}\in G$  are fixed. Since for ${x}_{1}^{n}\in {G}^{n}$  and ${x}_{1}^{n}={e}_{1}^{n}$  the value of ${P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)\left({x}_{1}^{n}\right)$  is uniquely determined, we verify only the case when ${x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)$  , $i=1,\dots ,n$  , for some ${y}_{1}^{s}\in {G}^{s}$  . If for some ${z}_{1}^{k}\in {G}^{k}$  and ${j}_{1},\dots ,{j}_{k}\in \left\{1,\dots ,n\right\}$  we have also ${x}_{i}={\mu }_{i}^{*}\left(\oplus {j}_{k}{j}_{1}{z}_{1}^{k}\right)$  , $i=1,\dots ,n$  , then ${\mu }_{i}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)={\mu }_{i}\left(\oplus {j}_{k}{j}_{1}{z}_{1}^{k}\right)$  for every $i=1,\dots ,n$  , which, according to (4 ), implies $g\oplus {i}_{s}{i}_{1}{y}_{1}^{s}=g\oplus {j}_{k}{j}_{1}{z}_{1}^{k}$  . This means that also in this case ${P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)\left({x}_{1}^{n}\right)$  is uniquely determined. Thus, the function ${P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)$  is single-valued.
Proposition 2. For all $g,{g}_{1},\dots ,{g}_{n},{h}_{1},{h}_{2}\in G$  we have ${P}_{\left({h}_{1},{h}_{2}\right)}\left(g\left[{g}_{1}^{n}\right]\right)={P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)\left[{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\dots {P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{n}\right)\right].$
• Proof. Let $g,{g}_{1},\dots ,{g}_{n}\in G$  and ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\left[{g}_{1}^{n}\right]\right)$  . If ${x}_{1}^{n}\in {G}^{n}$  , then ${h}_{1}⊏g\left[{g}_{1}^{n}\right]\left[{x}_{1}^{n}\right]\vee {h}_{2}⊏g\left[{g}_{1}^{n}\right]\left[{x}_{1}^{n}\right],$  whence, applying the superassociativity (3 ), we obtain  $\begin{array}{c}{h}_{1}⊏g\left[{g}_{1}\left[{x}_{1}^{n}\right]\dots {g}_{n}\left[{x}_{1}^{n}\right]\right]\vee {h}_{2}⊏g\left[{g}_{1}\left[{x}_{1}^{n}\right]\dots {g}_{n}\left[{x}_{1}^{n}\right]\right].\end{array}$ (17)
This together with the $v$  -negativity of $\chi$  implies  $\begin{array}{c}{h}_{1}⊏{g}_{i}\left[{x}_{1}^{n}\right]\vee {h}_{2}⊏{g}_{i}\left[{x}_{1}^{n}\right],i=1,\dots ,n.\end{array}$ (18)
From (17 ) it follows that $\left({g}_{1}\left[{x}_{1}^{n}\right],\dots ,{g}_{n}\left[{x}_{1}^{n}\right]\right)\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)$  , from (18 ) that ${x}_{1}^{n}\in {P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{i}\right)$  , $i=1,\dots ,n$  . So, if ${x}_{1}^{n}\in {G}^{n}$  , then  $\begin{array}{c}{x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\left[{g}_{1}^{n}\right]\right)⟷\left\{\begin{array}{c}\left({g}_{1}\left[{x}_{1}^{n}\right],\dots ,{g}_{n}\left[{x}_{1}^{n}\right]\right)\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right),\\ {\wedge }_{i=1}^{n}{x}_{1}^{n}\in {P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{i}\right).\end{array}\end{array}$ (19)
Analogously we can verify that  $\begin{array}{c}{e}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\left[{g}_{1}^{n}\right]\right)⟷\left\{\begin{array}{c}\left({g}_{1},\dots ,{g}_{n}\right)\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right),\\ {\wedge }_{i=1}^{n}{e}_{1}^{n}\in {P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{i}\right).\end{array}\end{array}$ (20)
Now let ${x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)$  , $i=1,\dots ,n$  , for some ${i}_{1},\dots ,{i}_{s}\in \left\{1,\dots ,n\right\}$  and ${y}_{1}^{s}\in {G}^{s}$  . Then ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\left[{g}_{1}^{n}\right]\right)$  implies ${h}_{1}⊏g\left[{g}_{1}^{n}\right]\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\vee {h}_{2}⊏g\left[{g}_{1}^{n}\right]\oplus {i}_{s}{i}_{1}{y}_{1}^{s},$  which, by (6 ), is equivalent to  $\begin{array}{c}{h}_{1}⊏g\left[{g}_{1}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\dots {g}_{n}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right]\vee {h}_{2}⊏g\left[{g}_{1}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\dots {g}_{n}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right].\end{array}$ (21)
From this, applying the $v$  -negativity of $\chi$  , we obtain  $\begin{array}{c}{h}_{1}⊏{g}_{i}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\vee {h}_{2}⊏{g}_{i}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\end{array}$ (22)
for every $i=1,\dots ,n$  .
The condition (21 ) is equivalent to $\left({g}_{1}\oplus {i}_{s}{i}_{1}{y}_{1}^{s},\dots ,{g}_{n}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)$  . The condition (22 ) shows that ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{i}\right)$  for every $i=1,\dots ,n$  , where ${x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)$  , $i=1,\dots ,n$  . So,  $\begin{array}{c}{x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\left[{g}_{1}^{n}\right]\right)⟷\left\{\begin{array}{c}\left({g}_{1}\oplus {i}_{s}{i}_{1}{y}_{1}^{s},\dots ,{g}_{n}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right),\\ {\wedge }_{i=1}^{n}{x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{i}\right),\end{array}\end{array}$ (23)
where ${x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)$  , $i=1,\dots ,n$  .
Let ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\left[{g}_{1}^{n}\right]\right)$  . If ${x}_{1}^{n}\in {G}^{n}$  , then, according to (16 ) and (19 ), we have  $\begin{array}{cc}{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\left[{g}_{1}^{n}\right]\right)\left({x}_{1}^{n}\right)& =g\left[{g}_{1}^{n}\right]\left[{x}_{1}^{n}\right]=g\left[{g}_{1}\left[{x}_{1}^{n}\right]\dots {g}_{n}\left[{x}_{1}^{n}\right]\right]\end{array}$
 $\begin{array}{cc}& ={P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)\left({g}_{1}\left[{x}_{1}^{n}\right],\dots ,{g}_{n}\left[{x}_{1}^{n}\right]\right)\end{array}$
 $\begin{array}{cc}& ={P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)\left({P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\left({x}_{1}^{n}\right),\dots ,{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{n}\right)\left({x}_{1}^{n}\right)\right)\end{array}$
 $\begin{array}{cc}& ={P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)\left[{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\dots {P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{n}\right)\right]\left({x}_{1}^{n}\right).\end{array}$
Similarly, we can prove that ${P}_{\left({h}_{1},{h}_{2}\right)}\left(g\left[{g}_{1}^{n}\right]\right)\left({e}_{1}^{n}\right)=\left[{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\dots {P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{n}\right)\right]\left({e}_{1}^{n}\right)$  for ${e}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\left[{g}_{1}^{n}\right]\right)$  . If ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\left[{g}_{1}^{n}\right]\right)$  , where ${x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)$  , $i=1,\dots ,n$  , for some ${y}_{1}^{s}\in {G}^{s}$  , ${i}_{1},\dots ,{i}_{s}\in \left\{1,\dots ,n\right\}$  , then, according to (16 ) and (23 ), we obtain  $\begin{array}{cc}{P}_{\left({h}_{1},{h}_{2}\right)}\left(g\left[{g}_{1}^{n}\right]\right)\left({x}_{1}^{n}\right)& =g\left[{g}_{1}^{n}\right]\oplus {i}_{s}{i}_{1}{y}_{1}^{s}=g\left[{g}_{1}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\dots {g}_{n}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right]\end{array}$
 $\begin{array}{cc}& ={P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)\left({g}_{1}\oplus {i}_{s}{i}_{1}{y}_{1}^{s},\dots ,{g}_{n}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)\end{array}$
 $\begin{array}{cc}& ={P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)\left({P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\left({x}_{1}^{n}\right),\dots ,{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{n}\right)\left({x}_{1}^{n}\right)\right)\end{array}$
 $\begin{array}{cc}& ={P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)\left[{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\dots {P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{n}\right)\right]\left({x}_{1}^{n}\right).\end{array}$
The proof is complete.
Proposition 3. For all ${g}_{1},{g}_{2},{h}_{1},{h}_{2}\in G$  and $i=1,\dots ,n$  we have ${P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\oplus i{g}_{2}\right)={P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\oplus i{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right).$
• Proof. Let ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\oplus i{g}_{2}\right)$  . If ${x}_{1}^{n}\in {G}^{n}$  , then ${h}_{1}⊏\left({g}_{1}\oplus i{g}_{2}\right)\left[{x}_{1}^{n}\right]\vee {h}_{2}⊏\left({g}_{1}\oplus i{g}_{2}\right)\left[{x}_{1}^{n}\right],$  which, by (5 ), is equivalent to  $\begin{array}{c}{h}_{1}⊏{g}_{1}\left[{x}_{1}^{i-1}{g}_{2}\left[{x}_{1}^{n}\right]{x}_{i+1}^{n}\right]\vee {h}_{2}⊏{g}_{1}\left[{x}_{1}^{i-1}{g}_{2}\left[{x}_{1}^{n}\right]{x}_{i+1}^{n}\right].\end{array}$ (24)
This, according to the $v$  -negativity of $\chi$  , implies  $\begin{array}{c}{h}_{1}⊏{g}_{2}\left[{x}_{1}^{n}\right]\vee {h}_{2}⊏{g}_{2}\left[{x}_{1}^{n}\right].\end{array}$ (25)
The condition (24 ) means that $\left({x}_{1}^{i-1},{g}_{2}\left[{x}_{1}^{n}\right],{x}_{i+1}^{n}\right)\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)$  . From (25 ) we obtain ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)$  . So, for ${x}_{1}^{n}\in {G}^{n}$  we have  $\begin{array}{c}{x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\oplus i{g}_{2}\right)⟷\left\{\begin{array}{c}\left({x}_{1}^{i-1},{g}_{2}\left[{x}_{1}^{n}\right],{x}_{i+1}^{n}\right)\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\\ {x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right).\end{array}\end{array}$ (26)
Consider now the case when ${x}_{1}^{n}={e}_{1}^{n}$  . In this case ${e}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\oplus i{g}_{2}\right)$  means, by (17 ), that  $\begin{array}{c}{h}_{1}⊏{g}_{1}\oplus i{g}_{2}\vee {h}_{2}⊏{g}_{1}\oplus i{g}_{2}.\end{array}$ (27)
Because ${g}_{1}\oplus i{g}_{2}⊏{\mu }_{i}\left(\oplus i{g}_{2}\right)={g}_{2}$  , by the $v$  -negativity of $\chi$  , the above condition gives  $\begin{array}{c}{h}_{1}⊏{g}_{2}\vee {h}_{2}⊏{g}_{2}.\end{array}$ (28)
But ${\mu }_{i}^{*}\left(\oplus i{g}_{2}\right)={\mu }_{i}\left(\oplus i{g}_{2}\right)={g}_{2}$  and ${\mu }_{k}^{*}\left(\oplus i{g}_{2}\right)={e}_{k}$  for $k\in \left\{1,\dots ,n\right\}\\left\{i\right\}$  , so, (27 ) implies $\left({e}_{1}^{i-1},{g}_{2},{e}_{i+1}^{n}\right)\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)$  . On the other hand, from (28 ) it follows ${e}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)$  . Therefore  $\begin{array}{c}{e}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\oplus i{g}_{2}\right)⟷\left\{\begin{array}{c}\left({e}_{1}^{i-1},{g}_{2},{e}_{i+1}^{n}\right)\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\\ {e}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right).\end{array}\end{array}$ (29)
In the third case when ${x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)$  , $i=1,\dots ,n$  , for some ${y}_{1}^{s}\in {G}^{s}$  , ${i}_{1},\dots ,{i}_{s}\in \left\{1,\dots ,n\right\}$  , from ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\oplus i{g}_{2}\right)$  we conclude  $\begin{array}{c}{h}_{1}⊏\left({g}_{1}\oplus i{g}_{2}\right)\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\vee {h}_{2}⊏\left({g}_{1}\oplus i{g}_{2}\right)\oplus {i}_{s}{i}_{1}{y}_{1}^{s}.\end{array}$ (30)
Since $\chi$  is $v$  -negative, we have $\left({g}_{1}\oplus i{g}_{2}\right)\oplus {i}_{s}{i}_{1}{y}_{1}^{s}⊏{\mu }_{i}\left(\oplus i{g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)={g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}$  , which means that (30 ) can be written in the form  $\begin{array}{c}{h}_{1}⊏{g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\vee {h}_{2}⊏{g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}.\end{array}$ (31)
But ${\mu }_{i}^{*}\left(\oplus i{g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)={\mu }_{i}\left(\oplus i{g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)={g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}$  and ${\mu }_{k}^{*}\left(\oplus i{g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)={\mu }_{k}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)$  for $k\in \left\{1,\dots ,n\right\}\\left\{i\right\}$  . This, together with the condition (30 ), proves $\left({x}_{1}^{i-1},{g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s},{x}_{i+1}^{n}\right)\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right).$  Similarly, from (31 ) we can deduce ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right).$  Therefore ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\oplus i{g}_{2}\right)⟷\left\{\begin{array}{c}\left({x}_{1}^{i-1},{g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s},{x}_{i+1}^{n}\right)\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\\ {x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right),\end{array}$  where ${x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)$  , $i=1,\dots ,n$  .
Let ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\oplus i{g}_{2}\right)$  . If ${x}_{1}^{n}\in {G}^{n}$  , then, according to (16 ) and (26 ), we have  $\begin{array}{cc}{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\oplus i{g}_{2}\right)\left({x}_{1}^{n}\right)& =\left({g}_{1}\oplus i{g}_{2}\right)\left[{x}_{1}^{n}\right]={g}_{1}\left[{x}_{1}^{i-1}{g}_{2}\left[{x}_{1}^{n}\right]{x}_{i+1}^{n}\right]\end{array}$
 $\begin{array}{cc}& ={P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\left({x}_{1}^{i-1},{g}_{2}\left[{x}_{1}^{n}\right],{x}_{i+1}^{n}\right)\end{array}$
 $\begin{array}{cc}& ={P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\left({x}_{1}^{i-1},{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)\left({x}_{1}^{n}\right),{x}_{i+1}^{n}\right)\end{array}$
 $\begin{array}{cc}& ={P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\oplus i{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)\left({x}_{1}^{n}\right).\end{array}$
If ${x}_{1}^{n}={e}_{1}^{n},$  then, analogously as in the previous case, using (16 ) and (29 ) we obtain ${P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\oplus i{g}_{2}\right)\left({e}_{1}^{n}\right)={P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\oplus i{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)\left({e}_{1}^{n}\right).$  Similarly, in the case when ${x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)$  , $i=1,\dots ,n$  , for some ${y}_{1}^{s}\in {G}^{s}$  , ${i}_{1},\dots ,{i}_{s}\in \left\{1,\dots ,n\right\}$  , we have  $\begin{array}{cc}{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\oplus i{g}_{2}\right)\left({x}_{1}^{n}\right)& =\left({g}_{1}\oplus i{g}_{2}\right)\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\end{array}$
 $\begin{array}{cc}& ={P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\left({x}_{1}^{i-1},{g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s},{x}_{i+1}^{n}\right)\end{array}$
 $\begin{array}{cc}& ={P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\left({x}_{1}^{i-1},{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)\left({x}_{1}^{n}\right),{x}_{i+1}^{n}\right)\end{array}$
 $\begin{array}{cc}& ={P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\oplus i{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)\left({x}_{1}^{n}\right).\end{array}$
This completes our proof.
Basing on these propositions we are able to prove the sufficiency of the conditions of Theorem 1 .
Sufficiency. Let the triplet $\left(\chi ,\gamma ,\pi \right)$  of binary relations on a representable Menger $\left(2,n\right)$  -semigroup $\mathcal{G}=\left(G;o,\oplus 1,\dots ,\oplus n\right)$  satisfies all the conditions of the theorem. Then, as it follows from Propositions 1 –3 , for all ${h}_{1},{h}_{2}\in G$  , the mapping ${P}_{\left({h}_{1},{h}_{2}\right)}$  is a representation of $\mathcal{G}$  by $n$  -place functions. Consider the family of representations ${P}_{\left({h}_{1},{h}_{2}\right)}$  such that $\left({h}_{1},{h}_{2}\right)\in \gamma$  . Let $P$  be the sum of this family, i.e. $P={\sum }_{\left({h}_{1},{h}_{2}\right)\in \gamma }{P}_{\left({h}_{1},{h}_{2}\right)}$  . Of course, $P$  is a representation of $\mathcal{G}$  by $n$  -place functions. Let us show that $\chi ={\chi }_{P}$  , $\gamma ={\gamma }_{P}$  and $\pi ={\pi }_{P}$  .
Let $\left({g}_{1},{g}_{2}\right)\in {\chi }_{P}$  . Then, according to (8 ), we have $\left({g}_{1},{g}_{2}\right)\in {\chi }_{\left({h}_{1},{h}_{2}\right)}$  $\text{1}$  for all $\left({h}_{1},{h}_{2}\right)\in \gamma$  , i.e. $\left(\forall \left({h}_{1},{h}_{2}\right)\in \gamma \right)\left(\text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\subset \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)\right),$  which is equivalent to $\left(\forall \left({h}_{1},{h}_{2}\right)\in \gamma \right)\left(\forall {x}_{1}^{n}\right)\left({x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)⟶{x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)\right).$  From this, for ${x}_{1}^{n}={e}_{1}^{n}$  , we obtain $\left(\forall \left({h}_{1},{h}_{2}\right)\in \gamma \right)\left({e}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)⟶{e}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)\right),$  which means that $\left(\forall \left({h}_{1},{h}_{2}\right)\in \gamma \right)\left({h}_{1}⊏{g}_{1}\vee {h}_{2}⊏{g}_{1}⟶{h}_{1}⊏{g}_{2}\vee {h}_{2}⊏{g}_{2}\right).$  Let ${g}_{1}\ne 0$  . Then ${g}_{1}\top {g}_{1}$  and the above implication gives ${g}_{1}⊏{g}_{1}⟶{g}_{1}⊏{g}_{2}$  . This proves $\left({g}_{1},{g}_{2}\right)\in \chi$  because $\chi$  is reflexive. If ${g}_{1}=0$  , then $0=0\left[{g}_{2}\dots {g}_{2}\right]⊏{g}_{2}$  , by the $v$  -negativity of $\chi$  . Hence $\left(0,{g}_{2}\right)\in \chi$  . So, $\left({g}_{1},{g}_{2}\right)\in \chi$  , i.e. ${\chi }_{P}\subset \chi$  .
Conversely, let $\left({g}_{1},{g}_{2}\right)\in \chi$  , $\left({h}_{1},{h}_{2}\right)\in \gamma$  and ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)$  . If ${x}_{1}^{n}\in {G}^{n}$  , then ${h}_{1}⊏{g}_{1}\left[{x}_{1}^{n}\right]\vee {h}_{2}⊏{g}_{1}\left[{x}_{1}^{n}\right]$  . Since the $l$  -regularity of $\chi$  together with ${g}_{1}⊏{g}_{2}$  implies ${g}_{1}\left[{x}_{1}^{n}\right]⊏{g}_{2}\left[{x}_{1}^{n}\right]$  , from the above we conclude ${h}_{1}⊏{g}_{2}\left[{x}_{1}^{n}\right]\vee {h}_{2}⊏{g}_{2}\left[{x}_{1}^{n}\right]$  , i.e. ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)$  . Similarly, in the case ${x}_{1}^{n}={e}_{1}^{n}$  , from ${e}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)$  it follows ${e}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)$  . In the case when ${x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)$  , $i=1,\dots ,n$  , for some ${y}_{1}^{s}\in {G}^{s}$  , ${i}_{1},\dots ,{i}_{s}\in \left\{1,\dots ,n\right\}$  , applying the $l$  -regularity of $\chi$  to ${g}_{1}⊏{g}_{2}$  , we obtain ${g}_{1}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}⊏{g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}$  , whence, in view of ${h}_{1}⊏{g}_{1}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\vee {h}_{2}⊏{g}_{1}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}$  , we obtain ${h}_{1}⊏{g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\vee {h}_{2}⊏{g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}$  . Therefore ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)$  , which proves $\text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\subset \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)$  for all $\left({h}_{1},{h}_{2}\right)\in \gamma$  . Thus $\left({g}_{1},{g}_{2}\right)\in {\chi }_{P}$  , i.e. $\chi \subset {\chi }_{P}$  . Consequently, $\chi ={\chi }_{P}$  . This, together with the condition $\left(c\right)$  formulated in the theorem, gives $\pi =\chi \cap {\chi }^{-1}={\chi }_{P}\cap {\chi }_{P}^{-1}={\pi }_{P}$  . So, $\pi ={\pi }_{P}$  .
Now let $\left({g}_{1},{g}_{2}\right)\in {\gamma }_{P}$  . Then, according to (8 ), we have $\left({g}_{1},{g}_{2}\right)\in {\gamma }_{\left({h}_{1},{h}_{2}\right)}$  for some $\left({h}_{1},{h}_{2}\right)\in \gamma$  , i.e. $\left(\exists \left({h}_{1},{h}_{2}\right)\in \gamma \right)\left(\text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\cap \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)\ne \mathbb{\varnothing }\right),$  which is equivalent to $\left(\exists \left({h}_{1},{h}_{2}\right)\in \gamma \right)\left(\exists {x}_{1}^{n}\right)\left({x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{1}\right)\wedge {x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}\left({g}_{2}\right)\right).$  This, for ${x}_{1}^{n}\in {G}^{n}$  implies ${h}_{1}⊏{g}_{1}\left[{x}_{1}^{n}\right]\vee {h}_{2}⊏{g}_{1}\left[{x}_{1}^{n}\right]$  and ${h}_{1}⊏{g}_{2}\left[{x}_{1}^{n}\right]\vee {h}_{2}⊏{g}_{2}\left[{x}_{1}^{n}\right]$  . From the above, in view of ${h}_{1}\top {h}_{2}$  and (15 ), we obtain ${g}_{1}\left[{x}_{1}^{n}\right]\top {g}_{2}\left[{x}_{1}^{n}\right]$  , whence, applying the $l$  -cancellativity of $\gamma$  , we get ${g}_{1}\top {g}_{2}$  , i.e. $\left({g}_{1},{g}_{2}\right)\in \gamma$  .
In the similar way, we can see that in the case ${x}_{1}^{n}={e}_{1}^{n}$  the condition $\left({g}_{1},{g}_{2}\right)\in \gamma$  also holds.
If ${x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)$  , $i=1,\dots ,n$  , for some ${y}_{1}^{s}\in {G}^{s},$  ${i}_{1},\dots ,{i}_{s}\in \left\{1,\dots ,n\right\}$  , then ${h}_{1}⊏{g}_{1}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\vee {h}_{2}⊏{g}_{1}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}$  and ${h}_{1}⊏{g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\vee {h}_{2}⊏{g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}$  , whence, by ${h}_{1}\top {h}_{2}$  and (15 ), we obtain ${g}_{1}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\top {g}_{2}\oplus {i}_{s}{i}_{1}{y}_{1}^{s}$  . This gives ${g}_{1}\top {g}_{2}$  because $\gamma$  is $l$  -cancellative. In this way we have proved that in any case ${\gamma }_{P}\subset \gamma$  .
Conversely, let $\left({g}_{1},{g}_{2}\right)\in \gamma$  . Since $\chi$  is reflexive, ${g}_{1}⊏{g}_{1}$  and ${g}_{2}⊏{g}_{2}$  , whence ${g}_{1}⊏{g}_{1}\vee {g}_{2}⊏{g}_{1}$  and ${g}_{1}⊏{g}_{2}\vee {g}_{2}⊏{g}_{2}$  . Consequently, ${e}_{1}^{n}\in \text{pr}1{P}_{\left({g}_{1},{g}_{2}\right)}\left({g}_{1}\right)$  and ${e}_{1}^{n}\in \text{pr}1{P}_{\left({g}_{1},{g}_{2}\right)}\left({g}_{2}\right)$  . Thus $\left({g}_{1},{g}_{2}\right)\in {\gamma }_{\left({g}_{1},{g}_{2}\right)}\subset {\gamma }_{P}$  , i.e.
$\gamma \subset {\gamma }_{P}$  . So, $\gamma ={\gamma }_{P}$  .
This completes the proof of the theorem. $\square$  Problem 1. Find the necessary and sufficient conditions under which the triplet $\left(\chi ,\gamma ,\pi \right)$  of binary relations will be faithful projection representable for a representable Menger $\left(2,n\right)$  -semigroup. Deleting from Theorem 1 the equality $\pi =\chi \cap {\chi }^{-1}$  we obtain the necessary and sufficient conditions under which the pair $\left(\chi ,\gamma \right)$  of binary relations is projection representable for a representable Menger $\left(2,n\right)$  -semigroup. Furthermore, all parts of the proof of this theorem connected with these two relations are valid. So, we have the following
Theorem 2. A pair $\left(\chi ,\pi \right)$  of binary relations on $G$  is projection representable for a representable Menger $\left(2,n\right)$  -semigroup $\mathcal{G}$  if and only if $\chi$  is an $l$  -regular and $v$  -negative quasi-order, $\gamma$  is an $l$  -cancellative $0$  -quasi-equivalence and the implication $\left(15\right)$  is satisfied.
Problem 2. Find the necessary and sufficient conditions under which the pair $\left(\chi ,\gamma \right)$  of binary relations will be faithful projection representable for a representable Menger $\left(2,n\right)$  -semigroup.
Let $\mathcal{G}=\left(G;o,\oplus 1,\dots ,\oplus n\right)$  be a representable Menger $\left(2,n\right)$  -semigroup.
Let us consider on $G$  the set ${T}_{n}\left(G\right)$  of mappings $t:x↦t\left(x\right)$  defined as follows:
• (a) $x\in {T}_{n}\left(G\right)$  , i.e. ${T}_{n}\left(G\right)$  contains the identity transformation of $G$  ,
• (b) if $i\in \left\{1,\dots ,n\right\}$  , $a,{b}_{1},\dots ,{b}_{i-1},{b}_{i+1},\dots ,{b}_{n}\in G$  and $t\left(x\right)\in {T}_{n}$  , then $a\left[{b}_{1}^{i-1}t\left(x\right){b}_{i+1}^{n}\right]\in {T}_{n}$  ,
• (c) ${T}_{n}$  contains those and only those mappings which are defined by (a) and (b).
Let us consider on $G$  two binary relations ${\delta }_{1}$  and ${\delta }_{2}$  defined in the following way:
• 1. $\left({g}_{1},{g}_{2}\right)\in {\delta }_{1}⟷{g}_{1}=t\left({g}_{2}\right)$  for some $t\in {T}_{n}$  ,
• 2. $\left({g}_{1},{g}_{2}\right)\in {\delta }_{2}⟷\left\{\begin{array}{c}{g}_{1}=\left(x\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)\left[\overline{z}\right]and{g}_{2}={\mu }_{i}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right)\left[\overline{z}\right]forsome\\ x\in G,{y}_{1}^{s}\in {G}^{s},\overline{z}\in {G}^{n},i,{i}_{1},\dots ,{i}_{s}\in \left\{1,\dots ,n\right\},\\ wherethesymbol\left[\overline{z}\right]canbeempty.\end{array}$
It is not difficult to see that ${\delta }_{1}$  and ${\delta }_{2}$  are $l$  -regular relations, additionally ${\delta }_{1}$  is a quasi-order. Moreover, a binary relation $\rho \subset G×G$  is $v$  -negative if and only if it contains ${\delta }_{1}$  and ${\delta }_{2}$  .
Let $\pi$  be an $l$  -regular equivalence on a representable Menger $\left(2,n\right)$  -semigroup $\mathcal{G}$  . Denote by $\chi \left(\pi \right)$  the binary relation ${f}_{t}\left({f}_{R}\left({\delta }_{2}\right)\circ {\delta }_{1}\circ \pi \right)$  , where ${f}_{R}$  and ${f}_{t}$  are respectively reflexive and transitive closure operations (cf. [3), and $\circ$  is a composition of relations, $\text{2}$  i.e.
 $\begin{array}{c}\chi \left(\pi \right)={f}_{t}\left({f}_{R}\left({\delta }_{2}\right)\circ {\delta }_{1}\circ \pi \right)={\bigcup }_{n=1}^{\infty }{\left(\left({\delta }_{2}\cup {△}_{G}\right)\circ {\delta }_{1}\circ \pi \right)}^{n}.\end{array}$ (32)
Since $\pi$  , ${\delta }_{1}$  and ${f}_{R}\left({\delta }_{2}\right)$  are reflexive $l$  -regular relations, $\chi \left(\pi \right)$  is an $l$  -regular quasi-order containing $\pi$  , ${\delta }_{1}$  and ${\delta }_{2}$  . So, $\chi \left(\pi \right)$  is a $v$  -negative quasi-order.
Proposition 4. $\chi \left(\pi \right)$  is the least $l$  -regular and $v$  -negative quasi-order containing $\pi$  .
• Proof. Let $\chi$  be an arbitrary $l$  -regular and $v$  -negative quasi-order containing $\pi$  . Then ${\delta }_{1}\subset \chi$  and ${\delta }_{2}\subset \chi$  , because $\chi$  is $v$  -negative. Thus, $\pi \subset \chi$  , ${\delta }_{1}\subset \chi$  and ${f}_{R}\left({\delta }_{2}\right)\subset \chi$  , whence ${f}_{R}\left({\delta }_{2}\right)\circ {\delta }_{1}\circ \pi \subset {\chi }^{3}\subset \chi$  . From this, applying the transitivity of $\chi$  , we obtain $\left({f}_{R}\left({\delta }_{2}\right)\circ {\delta }_{1}\circ \pi {\right)}^{n}\subset {\chi }^{n}\subset \chi$  for every natural $n$  .
Therefore ${\cup }_{n=1}^{\infty }{\left(\left({\delta }_{2}\cup {△}_{G}\right)\circ {\delta }_{1}\circ \pi \right)}^{n}\subset \chi$  , i.e. $\chi \left(\pi \right)\subset \chi$  .
Theorem 3. A pair $\left(\gamma ,\pi \right)$  of binary relations on a representable Menger $\left(2,n\right)$  -semigroup $\mathcal{G}$  is projection representable if and only if $\left(a\right)$  $\gamma$  is an $l$  -cancellative $0$  -quasi-equivalence, $\left(b\right)$  $\pi$  is an $l$  -regular equivalence such that $\chi \left(\pi \right)\cap \left(\chi \left(\pi \right){\right)}^{-1}\subset \pi$  , $\left(c\right)$  the following condition
 $\begin{array}{c}{h}_{1}\top {h}_{2}\wedge {h}_{1}{⊏}_{\pi }{g}_{1}\wedge {h}_{2}{⊏}_{\pi }{g}_{2}⟶{g}_{1}\top {g}_{2},\end{array}$ (33)
where $h{⊏}_{\pi }g$  means $\left(h,g\right)\in \chi \left(\pi \right)$  , is satisfied for all ${g}_{1},{g}_{2},{h}_{1},{h}_{2}\in G$  .
• Proof. Let $P$  be such representation on a representable Menger $\left(2,n\right)$  -semigroup $\mathcal{G}$  for which $\gamma ={\gamma }_{P}$  and $\pi ={\pi }_{P}$  . Then, by Proposition 3 , we have $\chi \left(\pi \right)\subset {\chi }_{P}$  , whence $\chi \left(\pi \right)\cap \left(\chi \left(\pi \right){\right)}^{-1}\subset {\chi }_{P}\cap {\chi }_{P}^{-1}={\pi }_{P}=\pi$  .
Assume now that the premise of (33 ) is satisfied. Then $\left({h}_{1},{h}_{2}\right)\in \gamma$  , $\left({h}_{1},{g}_{1}\right)\in \chi \left(\pi \right)$  and $\left({h}_{2},{g}_{2}\right)\in \chi \left(\pi \right)$  . Consequently, $\left({h}_{1},{h}_{2}\right)\in {\gamma }_{P}$  , $\left({h}_{1},{g}_{1}\right)\in {\chi }_{P}$  and $\left({h}_{2},{g}_{2}\right)\in {\chi }_{P}$  , i.e. $\text{pr}1P\left({h}_{1}\right)\cap \text{pr}1P\left({h}_{2}\right)\ne \mathbb{\varnothing }$  , $\text{pr}1P\left({h}_{1}\right)\subset \text{pr}1P\left({g}_{1}\right)$  and $\text{pr}1P\left({h}_{2}\right)\subset \text{pr}1P\left({g}_{2}\right)$  , whence $\text{pr}1P\left({g}_{1}\right)\cap \text{pr}1P\left({g}_{2}\right)\ne \mathbb{\varnothing }$  . So, $\left({g}_{1},{g}_{2}\right)\in {\gamma }_{P}=\gamma$  , which means that the condition (33 ) is valid. The necessity is proved. To prove the sufficiency, assume that the pair $\left(\gamma ,\pi \right)$  of binary relations satisfies all the conditions of the theorem and consider the triplet $\left(\chi \left(\pi \right),\gamma ,\pi \right)$  .
Then $\pi ={\pi }^{-1}\subset \left(\chi \left(\pi \right){\right)}^{-1}$  , because $\pi \subset \chi \left(\pi \right)$  . Therefore $\pi \subset \chi \left(\pi \right)\cap \left(\chi \left(\pi \right){\right)}^{-1}$  , which, together with the condition $\left(b\right)$  , gives $\pi =\chi \left(\pi \right)\cap \left(\chi \left(\pi \right){\right)}^{-1}$  .
This means that the triplet $\left(\chi \left(\pi \right),\gamma ,\pi \right)$  satisfies all the conditions of Theorem 1 . So, $\left(\chi \left(\pi \right),\gamma ,\pi \right)$  , and in the consequence, $\left(\gamma ,\pi \right)$  is projection representable. The sufficiency is proved.
Problem 3. Find the necessary and sufficient conditions under which the pair $\left(\gamma ,\pi \right)$  of binary relations will be faithful projection representable.
Applying the method of mathematical induction to (32 ) we can prove the following proposition.
Proposition 5. The condition $\left({g}_{1},{g}_{2}\right)\in \chi \left(\pi \right)$  , where ${g}_{1},{g}_{2}\in G$  , means that the system of conditions
 $\begin{array}{c}\begin{array}{c}{g}_{1}={x}_{0}\wedge {g}_{2}={x}_{n},\\ {\wedge }_{i=0}^{n-1}\left(\left(\begin{array}{c}{x}_{i}\equiv {t}_{i}\left(\left({y}_{i}\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}}\right)\left[{\overline{w}}_{i}\right]\right),\\ {x}_{i+1}={\mu }_{{k}_{i}}\left(\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}}\right)\left[{\overline{w}}_{i}\right]\end{array}\right)\vee {x}_{i}\equiv {t}_{i}\left({x}_{i+1}\right)\right)\end{array}\right\}\end{array}$ (34)
is valid for some $n\in \mathbb{N}$  , ${x}_{i},{y}_{i},{z}_{i}\in G$  , ${\overline{w}}_{i}\in {G}^{n}$  , ${t}_{i}\in {T}_{n}$  , ${k}_{i}\in \left\{1,\dots ,n\right\}$  .
In the sequel the formula ${\bigwedge }_{i=m}^{n}\left(\left(\begin{array}{c}{x}_{i}\equiv {t}_{i}\left(\left({y}_{i}\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}}\right)\left[{\overline{w}}_{i}\right]\right),\\ {x}_{i+1}={\mu }_{{k}_{i}}\left(\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}}\right)\left[{\overline{w}}_{i}\right]\end{array}\right)\vee {x}_{i}\equiv {t}_{i}\left({x}_{i+1}\right)\right)$  will be denoted by $\mathfrak{M}\left(m,n\right)$  .
The inclusion $\chi \left(\pi \right)\cap \left(\chi \left(\pi \right){\right)}^{-1}\subset \pi$  means that for all ${g}_{1},{g}_{2}\in G$  we have $\left({g}_{1},{g}_{2}\right)\in \chi \left(\pi \right)\wedge \left({g}_{2},{g}_{1}\right)\in \chi \left(\pi \right)⟶{g}_{1}\equiv {g}_{2},$  which, according to Proposition 5 , can be written as the system of conditions $\left({A}_{n,m}{\right)}_{n,m\in \mathbb{N}}$  , where ${A}_{n,m}:\mathfrak{M}\left(0,n-1\right)\wedge \mathfrak{M}\left(n+1,n+m\right)\wedge {x}_{0}={x}_{n+m}⟶{x}_{0}\equiv {x}_{n}.$  The system $\left({A}_{n,m}{\right)}_{n,m\in \mathbb{N}}$  is equivalent to the system $\left({A}_{n}{\right)}_{n\in \mathbb{N}}$  , where ${A}_{n}:\mathfrak{M}\left(0,n-1\right)\wedge {x}_{0}={x}_{n}⟶{x}_{0}\equiv {x}_{1}.$  Consider now the implication (33 ). According to (34 ) the condition $\left({h}_{1},{g}_{1}\right)\in \chi \left(\pi \right)$  means that
 $\begin{array}{c}{h}_{1}={x}_{0}\wedge \mathfrak{M}\left(0,n-1\right)\wedge {x}_{n}={g}_{1}\end{array}$ (35)
for some ${x}_{i},{y}_{i},{z}_{{k}_{i}},{t}_{i},{k}_{i},{\overline{w}}_{i}$  . Similarly, the condition $\left({h}_{2},{g}_{2}\right)\in \chi \left(\pi \right)$  means that
 $\begin{array}{c}{h}_{2}={x}_{n+1}\wedge \mathfrak{M}\left(n+1,n+m\right)\wedge {x}_{n+m+1}={g}_{2}\end{array}$ (36)
for some ${x}_{i},{y}_{i},{z}_{{k}_{i}},{t}_{i},{k}_{i},{\overline{w}}_{i}$  . So, (33 ) can be written as the system $\left({B}_{n,m}{\right)}_{n,m\in \mathbb{N}}$  of conditions ${B}_{n,m}:{x}_{0}\top {x}_{n+1}\wedge \mathfrak{M}\left(0,n-1\right)\wedge \mathfrak{M}\left(n+1,n+m\right)⟶{x}_{n}\top {x}_{n+m+1}.$  In this way we have proved
Theorem 4. A pair $\left(\gamma ,\pi \right)$  of binary relations on a representable Menger $\left(2,n\right)$  -semigroup $\mathcal{G}$  is projection representable if and only if $\left(a\right)$  $\gamma$  is an $l$  -cancellative $0$  -quasi-equivalence, $\left(b\right)$  $\pi$  is an $l$  -regular equivalence, $\left(c\right)$  the systems of conditions $\left({A}_{n}{\right)}_{n\in \mathbb{N}}$  and $\left({B}_{n,m}{\right)}_{n,m\in \mathbb{N}}$  are satisfied.
Theorem 5. A pair $\left(\chi ,\pi \right)$  of binary relations is $\left($  faithful $\right)$  projection representable for a representable Menger $\left(2,n\right)$  -semigroup $\mathcal{G}$  if and only if $\chi$  is an $l$  -regular and $v$  -negative quasi-order such that $\pi =\chi \cap {\chi }^{-1}$  .
• Proof. The necessity of these conditions follows from the proof of Theorem 1 .
To prove their sufficiency, for every element $g\in G$  we define an $n$  -place function ${P}_{a}\left(g\right):{\mathfrak{A}}^{*}\to G$  , where $a\in G$  , putting  $\begin{array}{c}{P}_{a}\left(g\right)\left({x}_{1}^{n}\right)=\left\{\begin{array}{cc}g\left[{x}_{1}^{n}\right]& \text{if}a⊏g\left[{x}_{1}^{n}\right]\text{and}{x}_{1}^{n}\in {G}^{n},\\ g& \text{if}a⊏g\text{and}{x}_{1}^{n}={e}_{1}^{n},\\ g\oplus {i}_{s}{i}_{1}{y}_{1}^{s}& \text{if}a⊏g\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\text{and}{x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right),\\ & i=1,\dots ,n,\text{for some}{y}_{1}^{s}\in {G}^{s},\\ & {i}_{1},\dots ,{i}_{s}\in \left\{1,\dots ,n\right\}.\end{array}\end{array}$ (37)
Since, for ${h}_{1}={h}_{2}=a\in G,$  the function ${P}_{\left({h}_{1},{h}_{2}\right)}\left(g\right)$  defined by (16 ) coincides with the function ${P}_{a}\left(g\right)$  , from Propositions 1 – 3 it follows that the mapping ${P}_{a}:g↦{P}_{a}\left(g\right)$  is a representation of $\mathcal{G}$  by $n$  -place functions. Further, analogously as in the proof of Theorem 1 , we can prove that ${P}_{0}={\sum }_{a\in G}{P}_{a}$  is a representation of $\mathcal{G}$  for which $\chi ={\chi }_{{P}_{0}}$  and $\pi ={\pi }_{{P}_{0}}$  . So, the pair $\left(\chi ,\pi \right)$  is projection representable for $\mathcal{G}$  .
Let us show that $\left(\chi ,\pi \right)$  is faithful projection representable. In [8it is proved that each representable Menger $\left(2,n\right)$  -semigroup has a faithful representation by $n$  -place functions. Let $\Lambda$  be such representation. Then obviously ${\chi }_{\Lambda }=G×G$  and ${\pi }_{\Lambda }=G×G$  . Consider the representation $P=\Lambda +{P}_{0}$  . Since $\Lambda$  is a faithful representation, $P$  is also faithful. Moreover ${\chi }_{P}={\chi }_{\Lambda }\cap {\chi }_{{P}_{0}}=G×G\cap \chi =\chi$  and ${\pi }_{P}={\pi }_{\Lambda }\cap {\pi }_{{P}_{0}}=G×G\cap \pi =\pi$  . So, $\left(\chi ,\pi \right)$  is faithful projection representable for $\mathcal{G}$  .
In the same manner, using the construction (37 ), we can prove the following theorem.
Theorem 6. A binary relation $\chi$  is $\left($  faithful $\right)$  projection representable for a representable Menger $\left(2,n\right)$  -semigroup if and only if it is an $l$  -regular, $v$  -negative quasi-order.
Theorem 7. A binary relation $\pi$  is $\left($  faithful $\right)$  projection representable for a representable Menger $\left(2,n\right)$  -semigroup if and only if it is an $l$  -regular equivalence such that $\chi \left(\pi \right)\cap \left(\chi \left(\pi \right){\right)}^{-1}\subset \pi$  .
• Proof. Consider the pair $\left(\chi \left(\pi \right),\pi \right)$  of binary relations, where $\chi \left(\pi \right)$  is defined by (32 ). In a similar way, as in the proof of Theorem 3 , we can prove that this pair satisfies all the conditions of Theorem 5 , whence we conclude the validity of Theorem 7 .
Since, as it was showed above, the inclusion $\chi \left(\pi \right)\cap \left(\chi \left(\pi \right){\right)}^{-1}\subset \pi$  is equivalent to the system of conditions $\left({A}_{n}{\right)}_{n\in \mathbb{N}}$  , the last theorem can be rewritten in the form:
Theorem 8. A binary relation $\pi$  is $\left($  faithful $\right)$  projection representable for a representable Menger $\left(2,n\right)$  -semigroup if and only if it is an $l$  -regular equivalence and the system of conditions $\left({A}_{n}{\right)}_{n\in \mathbb{N}}$  is satisfied.
Consider on a Menger $\left(2,n\right)$  -semigroup $\mathcal{G}$  the binary relation ${\chi }_{0}$  defined in the following way:
 $\begin{array}{c}{\chi }_{0}={f}_{t}\left({f}_{R}\left({\delta }_{2}\right)\circ {\delta }_{1}\right)={\bigcup }_{n=1}^{\infty }{\left(\left({\delta }_{2}\cup {△}_{G}\right)\circ {\delta }_{1}\right)}^{n},\end{array}$ (38)
where ${f}_{t}$  and ${f}_{R}$  are reflexive and transitive closure operations.
Proposition 6. ${\chi }_{0}$  is the least $l$  -regular and $v$  -negative quasi-order on $\mathcal{G}$  .
The proof of this proposition is analogous to the proof of Proposition 3 .
Theorem 9. A binary relation $\gamma$  is projection representable for a representable Menger $\left(2,n\right)$  -semigroup if and only if it is an $l$  -cancellative $0$  -quasi-equivalence and the following implication
 $\begin{array}{c}{h}_{1}\top {h}_{2}\wedge {h}_{1}{⊏}_{0}{g}_{1}\wedge {h}_{2}{⊏}_{0}{g}_{2}⟶{g}_{1}\top {g}_{2}\end{array}$ (39)
is satisfied for all ${h}_{1},{h}_{2},{g}_{1},{g}_{2}\in G$  , where $h{⊏}_{0}g$  means $\left(h,g\right)\in {\chi }_{0}$  .
• Proof. The necessity of (39 ) can be proved analogous as the necessity of (33 ) in the proof of Theorem 3 . To prove the sufficiency we consider the pair $\left({\chi }_{0},\gamma \right)$  . By Proposition 6 , this pair satisfies all demands of Theorem 2 , whence we conclude the validity of Theorem 9 .
Problem 4. Find the necessary and sufficient conditions under which $\gamma$  will be faithful projection representable.
Basing on the formula (38 ) we can prove the following proposition:
Proposition 7. From $\left({g}_{1},{g}_{2}\right)\in {\chi }_{0}$  , where ${g}_{1},{g}_{2}\in G$  , it follows that the system of conditions ${g}_{1}={x}_{0}\wedge {g}_{2}={x}_{n}\wedge {\bigwedge }_{i=0}^{n-1}\left(\left(\begin{array}{c}{x}_{i}={t}_{i}\left(\left({y}_{i}\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}}\right)\left[{\overline{w}}_{i}\right]\right),\\ {x}_{i+1}={\mu }_{{k}_{i}}\left(\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}}\right)\left[{\overline{w}}_{i}\right]\end{array}\right)\vee {x}_{i}={t}_{i}\left({x}_{i+1}\right)\right)$  is valid for $n\in \mathbb{N}$  , ${x}_{i},{y}_{i},{z}_{i}\in G,$  ${\overline{w}}_{i}\in {G}^{n},$  ${t}_{i}\in {T}_{n}$  , ${k}_{i}\in \left\{1,\dots ,n\right\}$  .
Denoting by $\mathfrak{N}\left(m,n\right)$  the formula ${\bigwedge }_{i=m}^{n}\left(\left(\begin{array}{c}{x}_{i}={t}_{i}\left(\left({y}_{i}\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}}\right)\left[{\overline{w}}_{i}\right]\right),\\ {x}_{i+1}={\mu }_{{k}_{i}}\left(\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}}\right)\left[{\overline{w}}_{i}\right]\end{array}\right)\vee {x}_{i}={t}_{i}\left({x}_{i+1}\right)\right),$  and using the same argumentation as in the proof of Theorem 4 , we can prove that the implication (39 ) is equivalent to the system of conditions $\left({C}_{n,m}{\right)}_{n,m\in \mathbb{N}}$  , where ${C}_{n,m}:{x}_{0}\top {x}_{n+1}\wedge \mathfrak{N}\left(0,n-1\right)\wedge \mathfrak{N}\left(n+1,n+m\right)⟶{x}_{0}\top {x}_{n+m+1}.$  So, the following theorem is true:
Theorem 10. A binary relation $\gamma$  is projection representable for a representable Menger $\left(2,n\right)$  -semigroup if and only if it is an $l$  -cancellative $0$  -quasi-equivalence and the system of conditions $\left({C}_{n,m}{\right)}_{n,m\in \mathbb{N}}$  is satisfied.

$\text{1}$  ${\chi }_{\left({h}_{1},{h}_{2}\right)}$  denotes this quasi-order which corresponds to the representation ${P}_{\left({h}_{1},{h}_{2}\right)}$  . Analogously are defined ${\gamma }_{\left({h}_{1},{h}_{2}\right)}$  and ${\pi }_{\left({h}_{1},{h}_{2}\right)}$  .

$\text{2}$  Remind that $\sigma \circ \rho =\left\{\left(a,c\right)|\left(\exists b\right)\left(a,b\right)\in \rho \wedge \left(b,c\right)\in \sigma \right\}$  , ${f}_{R}\left(\rho \right)=\rho \cup {△}_{A}$  , ${f}_{t}\left(\rho \right)={\cup }_{n=1}^{\infty }{\rho }^{n}$  , where ${\rho }^{n}={\underbrace{\rho \circ \rho \circ \dots \circ \rho }}_{n}$  , $\rho ,\sigma$  — binary relations on $A$  and ${△}_{A}=\left\{\left(a,a\right)|a\in A\right\}$  .

4 Pro jection representable relations on $\left(2,n\right)$  -semigroups

Let $\chi$  , $\gamma$  and $\pi$  be three binary relations on a $\left(2,n\right)$  -semigroup $\left(G;\oplus 1,\dots ,\oplus n\right)$  . Similarly as in the case of Menger $\left(2,n\right)$  -semigroups we say that the triplet $\left(\chi ,\gamma ,\pi \right)$  is (faithful) projection representable for a $\left(2,n\right)$  -semigroup $\left(G;\oplus 1,\dots ,\oplus n\right)$  , if there exists such (faithful) representation $P$  of $\left(G;\oplus 1,\dots ,\oplus n\right)$  by $n$  -place functions for which $\chi ={\chi }_{P}$  , $\gamma ={\gamma }_{P}$  and $\pi ={\pi }_{P}$  . Analogously we define the projection representable pairs and separate relations.
It is not difficult to verify that our Theorem 1 formulated for representable Menger $\left(2,n\right)$  -semigroup is also valid for representable $\left(2,n\right)$  -semigroups. The proof of this version of Theorem 1 is analogous to the proof of the previous version, but in the proof of the sufficiency instead the representation $P$  we must consider the representation ${P}^{\bullet }$  , which is the sum of the family of representations $\left({P}_{\left({h}_{1},{h}_{2}\right)}^{\bullet }{\right)}_{\left({h}_{1},{h}_{2}\right)\in \gamma }$  , where for every $g\in G$  ${P}_{\left({h}_{1},{h}_{2}\right)}^{\bullet }\left(g\right):{\mathfrak{A}}_{0}^{*}\to G$  ( ${\mathfrak{A}}_{0}^{*}={\mathfrak{A}}_{0}\cup \left\{\left({e}_{1},\dots ,{e}_{n}\right)\right\}$  see page ) is a partial $n$  -place function such that ${x}_{1}^{n}\in \text{pr}1{P}_{\left({h}_{1},{h}_{2}\right)}^{\bullet }\left(g\right)⟷\left\{\begin{array}{cc}{h}_{1}⊏g\vee {h}_{2}⊏g& \text{if}{x}_{1}^{n}={e}_{1}^{n},\\ {h}_{1}⊏g\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\vee {h}_{2}⊏g\oplus {i}_{s}{i}_{1}{y}_{1}^{s}& \text{if}{x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right),\\ & i=1,\dots ,n,\text{for}\\ & \text{some}{y}_{1}^{s}\in {G}^{s}\text{and}\\ & {i}_{1}\dots ,{i}_{s}\in \left\{1,\dots ,n\right\}\end{array}$  and ${P}_{\left({h}_{1},{h}_{2}\right)}^{\bullet }\left(g\right)\left({x}_{1}^{n}\right)=\left\{\begin{array}{cc}g& \text{if}{x}_{1}^{n}={e}_{1}^{n},\\ g\oplus {i}_{s}{i}_{1}{y}_{1}^{s}& \text{if}{x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right),\\ & i=1,\dots ,n,\text{for some}{y}_{1}^{s}\in {G}^{s}\\ & \text{and}{i}_{1}\dots ,{i}_{s}\in \left\{1,\dots ,n\right\}.\end{array}$  Also Theorem 2 is valid for $\left(2,n\right)$  -semigroups. Moreover, problems analogous to Problem 1 and Problem 2 can be posed for $\left(2,n\right)$  -semigroups, too.
Theorem 3 will be valid for $\left(2,n\right)$  -semigroups if we replace the relation $\chi \left(\pi \right)$  by the relation
 $\begin{array}{c}{\chi }^{\bullet }\left(\pi \right)={f}_{t}\left({f}_{R}\left({\delta }_{2}\right)\circ \pi \right)={\bigcup }_{n=1}^{\infty }{\left(\left({\delta }_{2}\cup {△}_{G}\right)\circ \pi \right)}^{n},\end{array}$ (40)
i.e. if we delete ${\delta }_{1}$  from the formula (32 ). Proposition 5 for $\left(2,n\right)$  -semigroups has the following form:
Proposition 8. The condition $\left({g}_{1},{g}_{2}\right)\in {\chi }^{\bullet }\left(\pi \right)$  , where ${g}_{1},{g}_{2}\in G,$  means that the system of conditions ${g}_{1}={x}_{0}\wedge {g}_{2}={x}_{n}\wedge {\bigwedge }_{i=0}^{n-1}\left(\left(\begin{array}{c}{x}_{i}\equiv {y}_{i}\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}},\\ {x}_{i+1}={\mu }_{{k}_{i}}\left(\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}}\right)\end{array}\right)\vee {x}_{i}\equiv {x}_{i+1}\right)$  is valid for some $n\in \mathbb{N}$  , ${x}_{i},{y}_{i},{z}_{i}\in G,$  ${k}_{i}\in \left\{1,\dots ,n\right\}$  .
Denoting by $\mathfrak{X}\left(m,n\right)$  the formula ${\bigwedge }_{i=m}^{n}\left(\left(\begin{array}{c}{x}_{i}\equiv {y}_{i}\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}},\\ {x}_{i+1}={\mu }_{{k}_{i}}\left(\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}}\right)\end{array}\right)\vee {x}_{i}\equiv {x}_{i+1}\right)$  and using the same argumentation as in the proof of Theorem 4 , we can prove
Theorem 11. A pair $\left(\gamma ,\pi \right)$  of binary relations on a representable $\left(2,n\right)$  -semigroup is projection representable if and only if $\gamma$  is an $l$  -cancellative $0$  -quasi-equivalence, $\pi$  is an $l$  -regular equivalence, and the systems of conditions ${A}_{n}^{\bullet }$  and ${B}_{n,m}^{\bullet }$  , where
 $\begin{array}{ccc}& & {A}_{n}^{\bullet }:\mathfrak{X}\left(0,n-1\right)\wedge {x}_{0}={x}_{n}⟶{x}_{0}\equiv {x}_{1},\end{array}$
 $\begin{array}{ccc}& & {B}_{n,m}^{\bullet }:{x}_{0}\top {x}_{n+1}\wedge \mathfrak{X}\left(0,n-1\right)\wedge \mathfrak{X}\left(n+1,n+m\right)⟶{x}_{n}\top {x}_{n+m+1}\end{array}$
are satisfied.
Theorem 5 is valid for $\left(2,n\right)$  -semigroups too, but in the proof, the representation ${P}_{a}$  defined by (37 ), must be replaced by the representation ${P}_{a}^{\bullet }$  , where ${P}_{a}^{\bullet }\left(g\right)\left({x}_{1}^{n}\right)=\left\{\begin{array}{cc}g& \text{if}a⊏g\text{and}{x}_{1}={e}_{1}^{n},\\ g\oplus {i}_{s}{i}_{1}{y}_{1}^{s}& \text{if}a⊏g\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\text{and}{x}_{i}={\mu }_{i}^{*}\left(\oplus {i}_{s}{i}_{1}{y}_{1}^{s}\right),\\ & i=1,\dots ,n,\text{for some}{y}_{1}^{s}\in {G}^{s},\\ & \text{and}{i}_{1},\dots ,{i}_{s}\in \left\{1,\dots ,n\right\}.\end{array}$  For $\left(2,n\right)$  -semigroups Theorem 6 has the same form as for Menger $\left(2,n\right)$  -semigroup, in Theorem 7 the relation $\chi \left(\pi \right)$  must be replaced by ${\chi }^{\bullet }\left(\pi \right)$  , and in Theorem 8 instead of ${A}_{n}$  we must use ${A}_{n}^{\bullet }$  .
Further, using the same argumentation as in the proof of Proposition 4 we can prove that the relation ${\chi }_{0}^{\bullet }={f}_{t}\left({f}_{R}\left({\delta }_{2}\right)\right)={\bigcup }_{n=1}^{\infty }{\left({\delta }_{2}\cup {△}_{G}\right)}^{n},$  where ${f}_{t}$  and ${f}_{R}$  are reflexive and transitive closure operations, is the least $l$  -regular and $v$  -negative quasi-order on a given $\left(2,n\right)$  -semigroup. Using this relation, we can prove the analog of Theorem 10 for $\left(2,n\right)$  -semigroups. The analog of Problem 4 can be posed too.
Proposition 7 for $\left(2,n\right)$  -semigroups has the following form:
Proposition 9. The condition $\left({g}_{1},{g}_{2}\right)\in {\chi }_{0}^{\bullet }$  , where ${g}_{1},{g}_{2}\in G$  , means that the system of conditions ${g}_{1}={x}_{0}\wedge {g}_{2}={x}_{n}\wedge {\bigwedge }_{i=0}^{n-1}\left(\left(\begin{array}{c}{x}_{i}={y}_{i}\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}},\\ {x}_{i+1}={\mu }_{{k}_{i}}\left(\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}}\right)\end{array}\right)\vee {x}_{i}={x}_{i+1}\right)$  is valid for $n\in \mathbb{N}$  , ${x}_{i},{y}_{i},{z}_{i}\in G$  .
Further, denoting by $\mathfrak{B}\left(m,n\right)$  the formula ${\bigwedge }_{i=m}^{n}\left(\left(\begin{array}{c}{x}_{i}={y}_{i}\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}},\\ {x}_{i+1}={\mu }_{{k}_{i}}\left(\oplus {k}_{{s}_{i}}{k}_{{1}_{i}}{z}_{{1}_{i}}^{{s}_{i}}\right)\end{array}\right)\vee {x}_{i}={x}_{i+1}\right)$  and using the same argumentation as in the proof of Theorem 10 , we can prove
Theorem 12. A binary relation $\gamma$  is projection representable for a representable $\left(2,n\right)$  -semigroup if and only if it is an $l$  -cancellative $0$  -quasi-equivalence and the system of conditions $\left({C}_{n,m}^{\bullet }{\right)}_{n,m\in \mathbb{N}}$  , where ${C}_{n,m}^{\bullet }:{x}_{0}\top {x}_{n+1}\wedge \mathfrak{B}\left(0,n-1\right)\wedge \mathfrak{B}\left(n+1,n+m\right)⟶{x}_{0}\top {x}_{n+m+1}$  is satisfied.
References

1. Dudek W.A., Trokhimenko V.S. Functional Menger $\mathcal{P}$  -algebras, Commun. Algebra 30 (2002), $5921-5931.$
2. Mann H. On orthogonal Latin squares, Bull. Amer. Math. Soc. 50 (1944), 249-257.
3. Riguet J. Relations binaires, fermetures, correspondances de Galois, Bull. Soc. Math. France 76 (1948), $114-155$  .
4. Schein B.M. A relation of co-definability on semigroups of functions, (Russian), Ordered sets and lattices 1 (1971), $86-89,$  (Izdat. Saratov. Gos. Univ.)
5. Schein B.M. Projection partitions of function semigroups, C. R. Math. Rep. Acad. Sci. Canada 1 (1978/79), no. 2, $67-70.$
6. Schein B. M. Lectures on semigroups of transformations, Amer. Math. Soc. Translat. (2), 113 (1979), $123-181.$
7. Schein B. M., Trohimenko V. S. Algebras of multiplace functions, Semigroup Forum 17 (1979), $1-64.$
8. Sokhatsky F. N. An abstract characterization of $\left(2,n\right)$  -semigroups of $n$  -ary operations, (Russian), Mat. Issled. 65 (1982), $132-139.$
9. Trokhimenko V.S. Ordered algebras of multiplace functions, (Russian), Izv. Vyssh. Uchebn. Zaved. Matematika 1 (1971), $90-98.$
10. Trokhimenko V.S. Abstract characterizations of certain algebras of multiplace functions, (Russian), Izv. Yyssh. Uchebn. Zaved. Matematika 4 (1971), $87-95.$
11. Trokhimenko V.S. Characterization of the co-definability relation on ordered algebras of multiplace functions, (Russian), Izv. Vyssh. Uchebn. Zaved. Matematika 9 (1977), $80-88.$
12. Yakubov T. On $\left(2,n\right)$  -semigroups of $n$  -ary operations, (Russian), Bull. Akad. Ş tiinţ a SSR Moldov. 1 (1974), $29-46.$

Wiesław A. Dudek Institute of Mathematics, Technical University, 50-370 Wrocław, Poland E-mail: dudek@im.pwr.wroc.pl Valentin S. Trokhimenko Department of Mathematics, Pedagogical University, 21100 Vinnitsa, Ukraine E-mail: vtrokhim@sovamua.com