Does a statement equivalent to Theorem 3 below appear in the literature? If it does, what is the earliest published reference?

Theorem 1. Let $W$ be a non-trivial cancellative invertible-free [1] monoid, $A$ a non-empty set on which $W$ acts on the right, and $\cal{A}$ the semicategory with object set $A$ and morphisms triples $(a,w,b)\in A\times (W\backslash\{1\})\times A$ such that $b=a\cdot w$ with composition of morphisms $(a,w,b)\circ(b,v,c)=(a,wv,c)$ . Then every morphism of $\cal{A}$ is a bimorphism and no morphism is invertible.

Proof. Left as exercise.

Theorem 2. Let $W$ and $A$ be as above. The quotient of the free monoid in the morphisms of $\cal{A}$ by the congruence generated by the composition of morphisms, i.e. $(a,w,b)(b,v,c)\sim(a,wv,c)$, is cancellative and invertible-free.

Proof. Left as an exercise.

Denote the monoid in Theorem 2 as $\cal{A}^\circ$.

Theorem 3. Let $W$ and $A$ be as above. Then there exists a cancellative invertible-free monoid $X$, set $B$ of homographies from $W$ to $X$, and bijection $\varphi\colon A\leftrightarrow B$ that satisfies $\varphi(a\cdot w)=\varphi(a)^w$ for all $a\in A$ and $w\in W$.

Proof. Let $\cal{A}$ be as above and let $X=\cal{A}^\circ$. Let $\varphi_a(w)=[a,w,a\cdot w]$, the equivalence class that contains $(a,w,a\cdot w)$, if $w\not=1$ and $\varphi_a(1)=1$ for all $a$.

Note that if $W=\{1\}$, then $\varphi_a=\varphi_b$ for all $a,b\in A$.

[1] For definitions of these terms, see Can every cancellative invertible-free monoid be embedded in a group? and Is every invertible-free cancellative monoid action represented by "shifting" certain maps?

  • 1
    $\begingroup$ You probably want W nontrivial. I suspect that it is known the universal monoid of a cancelative senicategory is cancellative. I dont know what a bimorphism is. I think you are the first person looking at homographies. This seems a nice generalization of what I did for free monoids $\endgroup$ Jan 3, 2021 at 15:39
  • $\begingroup$ Hello, and good morning. Yes, the construction is similar if not exactly the same. Bimorphisms are defined here [1] as morphisms that satisfy left and right cancellation. $W$ is assumed non-trivial throughout except for the last sentence, which was intended to be subjunctive. Taking away the identity-preserving requirement, a homography is just a prefix non-decreasing [2] map. [1] en.wikipedia.org/wiki/Category_(mathematics) [2] See p. 501 of Nivat, M., Wirsing, M. (1996). Algebraic Methodology and Software Technology: 5th International Conference, AMAST '96 Munich, Germany, July 1996. $\endgroup$ Jan 3, 2021 at 16:01


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.