The Tutte-Grothendieck group of a convergent alphabetic rewriting system

The two operations, deletion and contraction of an edge, on multigraphs directly lead to the Tutte polynomial which satisfies a universal problem. As observed by Brylawski in terms of order relations, these operations may be interpreted as a particular instance of a general theory which involves universal invariants like the Tutte polynomial, and a universal group, called the Tutte-Grothendieck group. In this contribution, Brylawski's theory is extended in two ways: first of all, the order relation is replaced by a string rewriting system, and secondly, commutativity by partial commutations (that permits a kind of interpolation between non commutativity and full commutativity). This allows us to clarify the relations between the semigroup subject to rewriting and the Tutte-Grothendieck group: the later is actually the Grothendieck group completion of the former, up to the free adjunction of a unit (this was even not mention by Brylawski), and normal forms may be seen as universal invariants. Moreover we prove that such universal constructions are also possible in case of a non convergent rewriting system, outside the scope of Brylawski's work.


Introduction
In his paper [19], Tutte took advantage of two natural operations on (finite multi)graphs (actually on isomorphism classes of multigraphs), deletion and contraction of an edge, in order to introduce the ring Z[x, y] and a polynomial in two commuting variables x, y, also known by Whitney [22], unique up to isomorphism since solutions of a universal problem. This polynomial, since called the Tutte polynomial, is a graph invariant in at least two different meanings: first of all, it is defined on isomorphism classes, rather than on actual graphs, in such a way that two graphs with distinct Tutte polynomials are not isomorphic (a well-known functorial point of view), and, secondly, it is invariant with respect to a graph decomposition. Indeed, let G be a graph, and let e be an edge of G, which is not a loop (an edge with the same vertex as source and target) nor a bridge (an edge that connects two connected components of a graph). The edge contraction G/e of G is the graph obtained by identifying the vertices source and target of e, and removing the edge e. We write G − e for the graph where the edge e is merely removed; this operation is the edge deletion. Let us consider the graph G/e + (G − e) (well-defined as isomorphic classes) which can be interpreted as a decomposition of G. Then, the Tutte polynomial t is invariant with respect to this decomposition in the sense that t(G) = t(G/e) + t(G − e). Moreover this decomposition eventually terminates with graphs with bridges and loops only as edges, and the choice of edges to decompose is irrelevant.
In his paper [8], Brylawski observed that the previous construction (and many others, for instance the Tutte polynomial for matroids) may be explained in terms of an elegant and unified categorical framework (namely a universal problem of invariants). In brief, Brylawski considered an abstract notion of decomposition. Let X be a set, and let ≤ be an order relation on (a part of) the free commutative semigroup X ⊕ (actually Brylawski considered multisets, nevertheless the choice is here made to deal with semigroups since they play a central rôle in this contribution), which satisfies a certain number of axioms that are quickly reviewed in informal terms below for the sake of completeness (Appendix A contains a short review of Brylawski's theory in mathematical terms but it may be skipped) and to show how natural is their translations in terms of rewriting systems.
Let D(X) be a set of formal (finite) sums 1≤i≤k n i x i where x i ∈ X, n i ∈ N not all of them being zero (an element of the free commutative semigroup X ⊕ on X) partially ordered by ≤. If f, g ∈ D(X) such that f ≤ g, then we say that f decomposes into g or that g is a decomposition of f . Therefore D(X) is seen as a set of commutative decompositions. Elements of X that belong to D(X) are assumed to be minimal with respect to ≤. Elements of X ∩ D(X) that are maximal (and therefore incomparable since also minimal) are said to be irreducible. According to a second axiom satisfied by the order relation ≤, an element f ∈ D(X) cannot be decomposed further into any other element of D(X) if, and only if, f is a finite linear combination, with non negative integers as coefficients, of incomparable elements, that is, if Irr(X) is the set of all irreducibles, then f is not decomposed into another element if, and only if, f is a formal (finite) sum of elements of Irr(X) with non negative integers as coefficients. This property is similar to the notion of termination in rewriting systems. Two other properties (refinability and finiteness) on D(X) ensure that every element of X has one, and only one, "terminal" decomposition into irreducible elements. They are equivalent to convergence of a rewriting system. For instance, the order G < (G/e) + (G − e) on the free commutative semigroup generated by all (isomorphism classes of) finite graphs satisfies these axioms and properties. Now, to a decomposition (D(X), ≤) with the above properties may be attached a group in a universal way. A function f from X to an Abelian group G is said to be invariant if for every x ∈ X such that x ≤ 1≤i≤k n i x i (x i ∈ X, and n i ∈ N), then f (x) = n i f (x i ). Recall here that Tutte polynomial t is invariant because t(G) = t(G/e) + t(G − e). Brylawski proved the following theorem, which was his main result. There exist an Abelian group, called Tutte-Grothendieck group, and an invariant mapping t : X → A, called universal Tutte-Grothendieck invariant, such that for every Abelian group G and every invariant mapping f : X → G, there exists a unique group homorphism h : A → G with h • t = f . In addition, A is isomorphic to the free Abelian group with the irreducible elements as generators. In the classical context of graph theory, as expected, A is the additive structure of Z[x, y] and t is the Tutte polynomial. Many other decompositions enter in the scope of Brylawski's theory (see his paper [8], examples and references therein).
In the present contribution, we adapt Brylawski's results to the theory of (string) rewriting systems which we think is the natural framework to deal with theoretical notions of decomposition. Moreover we extend previous works by allowing non commutative, and even partially commutative, decompositions. Our main result, theorem 15, similar to Brylawski's main theorem, states the existence and uniqueness of a universal group and a universal invariant associated to some kind of string rewriting systems, even if there are not convergent (which is beyond the scope of Brylawski's work). In case of convergence, we prove that the universal group under consideration is the free partially commutative group generated by irreducible letters, which is a generalization of the original result, and that the universal invariant is nothing else than the normal form function that maps an element to its normal form. We mention the fact that in this case, the universal group is proved to be the Grothendieck completion of a monoid (obtained from the semigroup subject to rewriting by free adjunction of an identity), which was not seen by Brylawski even if he called Tutte-Grothendieck his universal construction.

Some universal constructions
The categorical notions used in this contribution, that are not defined here, come from [14]. This section is devoted to the presentation of Grothendieck group completion and free partially commutative structures which are used here after.

Basic notions and some notations
In what follows S , M and G denote the well-known categories of (small 1 ) semigroups, monoids and groups respectively, with their usual arrows (the so-called homomorphisms of semigroups, monoids or groups). Each of the categories S , M and G has a free object freely generated by a given (small) set. In other terms their forgetful functors to the category of sets have a left adjoint. In what follows we denote by X + , X * and F (X) respectively the free semigroup, monoid, group generated by X (see [6]), and we identify X as a subset of each of these algebraic structures. Note also that we denote by X ⊕ the free commutative semigroup on X.
There are also obvious forgetful functors from G to M , and from M to S (therefore also from G to S by composition). Both of them have a left adjoint (see [14]). The left adjoint of the forgetful functor from M to S is known to be the free adjunction S 1 = S ⊔ { 1 } of a unit to a semigroup S in order to obtain a monoid in a natural way (the symbol "⊔" denotes the set-theoretical disjoint sum). The unit of this adjunction, i S ,S : x ∈ S → x ∈ S 1 , which is an homomorphism of semigroups, is obviously into.
The forgetful functor from G → M has both a left and a right adjoint. Its right adjoint is given, at the object level, as a class mapping that associates a monoid to its group of invertible elements. Its left adjoint, more involved, is described below as group completion.

Group completion
The left adjoint of the forgetful functor from groups to monoids may be described as the (unique) solution of the following universal problem. Let M be a monoid. Then there exists a unique group G(M ), called the group completion or universal enveloping group or Grothendieck group of M (see [21] and references therein, and also [15]), and a unique homomorphism of monoids i M ,M : M → G(M ) such that for every group G and every homomorphism of monoids f : M → G, there is a unique homomorphism of groups f : G(M ) → M such that the following diagram commutes (in the category of monoids).
It is not difficult to check that G(M ) is given either as F (M ) / I M where I M is the subset { mn(m * n) −1 : m, n ∈ M } (where " * " is the monoid multiplication of M , and where F (X) denotes the free group of X, see Subsection 2.1 and if G is a group and A is any subset of G, then A is the normal subgroup of G generated by A), see [15], or as the quotient : m ∈ M } (here the star " * " stands for the free monoid functor, see also Subsection 2.1, and ǫ is the empty word) where M −1 is the set of (formal) symbols { m −1 : m ∈ M } equipotent to M .

Free partially commutative structures
Other universal problems, which will play an important rôle in what follows, are the free partially commutative structures. These structures have been introduced in [9] (see also [20]). A good review of these objects is [12]. Since such constructions may be performed in any of the categories of semigroups, monoids and groups, they are presented here in a generic way on a category C ∈ { S , M , G } so that all statements make sense in any of these categories. Let X be a set and let θ ⊆ X × X be a symmetric (i.e., for every x, y ∈ X, (x, y) ∈ θ implies (y, x) ∈ θ) and reflexive relation on X (i.e., for each x ∈ X, (x, x) ∈ θ). Let C be an object in C , and f : X → C be a set-theoretical mapping. This function is said to respect the commutations whenever ( It can be shown that there exists a unique object C (X, θ) of C and a unique mapping j C ,X : X → C(X, θ) that respects the commutations such that for every object C of C and every mapping f : X → C that respects the commutations, there is a unique arrow (in C ) f C : C (X, θ) → C such that the following diagram commutes in the category of sets.
The object C (X, θ) is usually called the free partially commutative semigroup (respectively, monoid, group) on X (or on (X, θ) to be more precise) depending on C , and may be constructed as follows: We may note that C (X, ∅) is nothing else than the usual free (non commutative) object in the category C , while C (X, (X × X) \ ∆ X ), where ∆ X is the equality relation on X, is the free commutative object in C (in particular, S (X, (X × X) \ ∆ X ) = X ⊕ is the free commutative semigroup).
We may clarify the relations between the free partially commutative structures. Using universal properties, it is not difficult to check that M (X, θ) is isomorphic to S (X, θ) 1 (actu- where ǫ is the empty word) in such a way that S (X, θ) embeds in M (X, θ) as a sub-semigroup.
◮ Lemma 1. The monoid M (X, θ) is isomorphic to the free adjunction S (X, θ) 1 of an identity to the semigroup S (X, θ).
Proof. To prove this lemma it is sufficient to check that M (X, θ) is a solution of the universal problem of adjunction of a unit to S (X, θ). According to the universal problem of the free partially commutative semigroup S (X, θ), there is a unique homomorphism of semigroups I : S (X, θ) → M (X, θ) such that the following diagram is commutative.
Now, let M be a monoid and f : S (X, θ) → M be a semigroup homomorphism. Therefore there exists f 0 : X → M that respects the commutations and such that The relations between all the arrows are summarized in the following commutative diagram.
There is also an important relation between G (X, θ) and M (X, θ) given in the following lemma.
Proof. The set-theoretical mapping j G ,X : X → G (X, θ) respects the commutations, therefore according to the universal problem of the free partially commutative monoid over (X, θ) there is a unique homomorphism of monoids j M G ,X that makes commute the following diagram.
Now, let G be any group, and f : M (X, θ) → G be an homomorphism of monoids. Then, according to the universal problem of the free partially commutative monoid, there is a unique set-theoretical mapping f 0 : X → G that respects the commutations and f • j M ,X = f 0 . Now The relations between all the arrows are summarized in the following commutative diagram.
◭ Actually a result from [11] page 66 (see also [12]) states that the natural mapping j M G ,X of the proof of lemma 2 is one-to-one so that M (X, θ) may be identified with a sub-monoid of its Grothendieck completion G (X, θ).
◮ Definition 3. Let X be any set. For every x ∈ X and every w ∈ X * , let us define |w| x as the number of occurrences of the letter x in the word w. More precisely, if ǫ is the empty word, then |ǫ| x = 0, |y| x = 0 if y = x, |y| x = 1 if y = x for all y ∈ X, and if the length of w ∈ X * is > 1, then w = yw ′ for some letter y ∈ X, and w ′ ∈ X + , then |w| x = |y| x + |w ′ | x . Let ≡ be a congruence on X + or X * . It is said to be multi-homogeneous if for every w, w ′ in X + or X * , such that w ≡ w ′ , then for every x ∈ X, |w| x = |w ′ | x . Therefore we may define |[w] ≡ | x = |w| x for the class [w] ≡ of w modulo ≡ (it does not depend on the representative of the class modulo ≡).
According to [12], any congruence of the form ≡ θ is a multi-homogenous congruence, so that we may define |w| x for all w ∈ C (X, θ) and all x ∈ X (where C = S or M ). The notion of multi-homogeneity is used to check that we may identify the alphabet X has a generating set of C (X, θ) using the map j C ,X , which is shown to be into, in such a way that we consider that X ⊆ C (X, θ). Indeed, for semigroup or monoid case, let x, y ∈ X such that their classes modulo ≡ θ be equal. But ≡ θ is a multi-homogenous congruence (see [12]). Therefore x = y. Concerning the group case, let us assume that x, y ∈ X are equivalent modulo the normal subgroup Because the group is free, it means that x = y (no non trivial relations between the generators). In the sequel, we will treat X as a subset of C (X, θ).
More generally, let (X, θ) be a commutation alphabet and let Y ⊆ X.
as illustrated in the following lemma.

◮ Lemma 4. Under the previous assumptions, there is an arrow
Proof. Let incl : Y → X be the canonical inclusion. Define J : C (Y, θ Y ) → C (X, θ) as the unique arrow (in C ) such that the following diagram commutes.
Let w 0 ∈ C (Y, θ Y ). Let us define π w0 : X → Y such that π w0 (y) = y for every y ∈ Y ⊆ X, and π w0 (x) = w 0 for x ∈ X \ Y . We note that π w0 • incl = id Y . Then we may consider Π w0 : C (X, θ) → C (Y, θ Y ) as the unique arrow (in C ) that makes commute the following diagram. θY ) , and then J is into (and Π w0 is onto). ◭ According to lemma 4 we identify C (Y, θ Y ) as a sub-semigroup, sub-monoid or sub-group (depending on the choice of C ) of C (X, θ). In such situations we may use the following characterization.
◮ Lemma 5. Let (X, θ) be a commutation alphabet, and let Y ⊆ X be any subset. Let w ∈ S (X, θ). The following statements are equivalent: Because ≡ θ is a multi-homogeneous congruence, |ω| x = |w| x for all ω ∈ w and x ∈ X. Then the point 2. is obtained. Now, let w ∈ S (X, θ) such that for all x ∈ X, |w| x = 0 implies that x ∈ Y . Then, for all ω ∈ w (ω ∈ X + ), |ω| x = 0 for all x ∈ Y which means that ω ∈ Y + , and therefore w ∈ S (Y, θ Y ) so that 1. is obtained. ◭

Abstract rewriting systems
In this short section, as in the following, we adopt several notations and definitions from [1] that we summarize here. Let E be a set, and ⇒⊆ E × E be any binary relation, called a (one-step) reduction relation, and (E, ⇒) is called an abstract rewriting system. We denote by "x ⇒ y" the membership "(x, y) ∈⇒", and "x ⇒ y" stands for "(x, y) ∈⇒". Let R * be the reflexive transitive closure of a binary relation R. We use x ⇐ y or x * ⇐ = y to mean that y ⇒ x or y * = ⇒ x. An element x ∈ E is said to be reducible if there exists y ∈ E such that x ⇒ y. x is irreducible if it is not reducible, or, in other terms, if x is ⇒-minimal: there is no y ∈ E such that x ⇒ y. A normal form of x is an irreducible element y ∈ E such that x * = ⇒ y. If it exists, the normal form of x is denoted by N (x). The set of all normal forms, or equivalently, of all irreducible elements is denoted by Irr(E, ⇒) or Irr(E) when this causes no ambiguity. Note that two distinct normal forms x, y are ⇒-incomparable, that is x ⇒ y and y ⇒ x. A reduction relation ⇒ is said to be terminating or Noetherian if there is no infinite ⇒-descending chain (x n ) n∈N of elements of E such that x n ⇒ x n+1 for every n ≥ 0. In particular, if ⇒ is terminating, then it is irreflexive (otherwise x n = x for some x ∈ E such that x ⇒ x ∈ R would be an infinite ⇒-descending chain), that is the reason why we freely make use terminology from order relations (such as minimal, Noetherian, descending chain, etc.). We also say that the abstract rewriting system (E, →) is terminating or Noetherian whenever ⇒ is so. Two elements x, y ∈ E are said to be joignable if there is some z ∈ E such that x * = ⇒ z * ⇐ = y, and ⇒ (and also (E, ⇒)) is said to be confluent if for every x, y 1 , y 2 ∈ E such that y 1 * ⇐ = x * = ⇒ y 2 , then y 1 , y 2 are joignable. A reduction relation ⇒, and an abstract rewriting system (E, ⇒), are said to be convergent if it they are both confluent and terminating. Such reduction relations are interesting because in this case any element of E has one, and only one, normal form, and if we denote by * ⇐ ⇒ the reflexive transitive symmetric closure of ⇒ (that is the least equivalence relation on E containing ⇒), then x * ⇐ ⇒ y if, and only if, N (x) = N (y), therefore N : E → Irr(E) satisfies N (N (x)) = N (x) and so is onto and moreover, the function N : E / * ⇐ ⇒ → Irr(E) which maps the class of x modulo * ⇐ ⇒ to N (x) is well-defined, onto and one-to-one.

Semigroup rewriting systems
Now, let us assume that E is actually a semigroup S. Let R ⊆ S × S be any binary relation. We define the following relation ⇒ R ⊆ S × S by x ⇒ R y if, and only if, there are u, v ∈ S 1 and (a, b) ∈ R such that x = uav and y = ubv. A relation ⇒ R is called the (one-step) reduction rule associated with R. A relation R ⊆ S × S is said to be two-sided compatible if (x, y) ∈ R (x, y ∈ S) implies (uxv, uyv) ∈ R. Now, the intersection of the family of all two-sided compatible relations containing a given R ⊆ S × S (this family is non void since it contains the universal relation S × S) also is a two-sided compatible relation, and so we obtain the least two-sided compatible relation that contains R. It is called the two-sided compatible relation generated by R, and it can be shown that this is precisely ⇒ R . Now, given R ⊆ S ×S, (S, ⇒ R ) is called a (semigroup) rewriting system; definitions and properties of an abstract rewriting system may be applied to such a rewriting system. When S is the free monoid X * , then this kind of rewriting systems are known as string rewriting systems or semi-Thue systems (see [5]). We note that the reflexive transitive symmetric closure * ⇐ ⇒ R of ⇒ R is actually a semigroup congruence, because ⇒ R is two-sided compatible. The quotient semigroup S / * ⇐ ⇒ R is called the Thue semigroup associated with the semigroup rewriting system (S, ⇒ R ).

4
The Tutte-Grothendieck group of a convergent alphabetic rewriting system

A free partially commutative structure on normal forms
◮ Definition 6. Let (X, θ) be a commutation alphabet, and R ⊆ X × S (X, θ). Then (S (X, θ), ⇒ R ) is called an alphabetic semigroup rewriting system.
From now on in this current subsection we assume that (S (X, θ), ⇒ R ) is a convergent alphabetic semigroup rewriting system.
We study some algebraic consequences of convergence of this alphabetic rewriting system on irreducible elements in the form of some lemmas and corollaries. The main result (proposition 11) of this subsection is that the set of all normal forms of a convergent alphabetic semigroup rewriting system is actually the free partially commutative semigroup in a canonical way, generated by the irreducible letters. ◮ Lemma 7. Let w, w ′ ∈ Irr(S (X, θ), ⇒ R ). Then, ww ′ ∈ Irr(S (X, θ), ⇒ R ). As a result, Proof. Let us assume that ww ′ ∈ Irr (S (X, θ), θ). But in this case, either w or w ′ is reducible, which is a contradiction. As a result, Irr(S (X, θ), ⇒ R ) ⊆ S (X, θ) is closed under the operation of S (X, θ) so that and N is an homomorphism of semigroups. It is obviously onto. ◭ ◮ Corollary 9. The semigroups Irr(S (X, θ), ⇒ R ) and S (X, θ) / * ⇐ ⇒ are isomorphic.
Proof. As introduced in Subsection 3.1, let N : S (X, θ) / * ⇐ ⇒ → Irr(S (X, θ), ⇒ R ) be the function that maps the class of w ∈ S (X, θ) modulo * ⇐ ⇒ to the normal form N (w). It is a one-to-one and onto set-theoretical mapping. But according to corollary 8, N is a semigroup homomorphism, in such a way that N also is. ◭ The fact that the rewriting system is alphabetic (Definition 6) actually implies that the (isomorphic) semigroups Irr(S (X, θ), ⇒ R ) and S (X, θ) / * ⇐ ⇒ are actually free partially commutative. The objective is now to prove this statement. In order to do that, we exhibit the commutation alphabet that generates them. Let Irr(X) = Irr(S (X, θ), ⇒ R ) ∩ X (recall from Subsection 2.3 that X is considered as a subset of S (X, θ)). It is clear that Indeed, for every x ∈ X, w ∈ S (X, θ), x ⇒ R w if, and only if, there are u, v ∈ M (X, θ), x 1 ∈ X, w 1 ∈ S (X, θ) such that x = ux 1 v and w = uw 1 v. Since ≡ θ is a multi-homogenous congruence (see subsection 2.3), u = v is the empty word, and This characterization of Irr(X) is used in the following lemma.
Proof. Let us assume that X = ∅ and Irr(X) = ∅. Let x ∈ X. Since x ∈ Irr(X), there is some w ∈ S (X, θ) such that (x, w) ∈ R. Because w ∈ S (X, θ), and X generates S (X, θ), it can be written as x 1 u for some x 1 ∈ X, and u ∈ M (X, θ). Because Replacing w by v 1 u, we may construct an infinite descending chain x ⇒ R w ⇒ R v 1 u ⇒ R · · · , which is impossible since ⇒ R is assumed to be convergent, and therefore terminating. So Irr(X) = ∅. ◭ ◮ Remark. Forthcoming proposition 11, lemmas 13 and 14, and theorem 15 are obviously valid when X = ∅.
The following lemma reveals the structure of free partially commutative semigroup of Irr(S (X, θ), ⇒ R ), and therefore also of S (X, θ) / * ⇐ ⇒ according to lemma 9.

The Tutte-Grothendieck group of a convergent alphabetic rewriting system
◮ Definition 12. Let (X, θ) be a commutation alphabet, and let ⇒ R be an alphabetic rewriting system. Let S be any semigroup, and let f : X → S that respects the commutations. Let f S : S (X, θ) → S be the unique homomorphism of semigroups such that the following diagram commutes (see Subsection 2.3).
Then f is said to be an R-invariant if for every x ∈ X and w ∈ S (X, θ) such that (x, w) ∈ R, then f (x) = f S (w).
Informally speaking, according to definition 12, a function f that respects the commutations is an R-invariant if its canonical semigroup extension f S is constant for all reductions (x, w) ∈ R.
Let us assume that (X, θ) is a commutation alphabet, and let ⇒ S be an alphabetic rewriting system on S (X, θ) (not necessarily convergent). The fact that the rewriting system is alphabetic implies in an essential way the following results.
◮ Lemma 13. Let S be a semigroup, and let f : X → S be a function that respects the commutations. Let f S be its canonical semigroup extension from S (X, θ) to S. If f is a R-invariant, then for every w, w ′ ∈ S (X, θ) such that w ⇒ R w ′ , we have f S (w) = f S (w ′ ).
Proof. Since we will deal with the empty word, one needs to recall the following. According to lemma 1, M (X, θ) = S (X, θ) ∪ { ǫ }, where ǫ is the empty word. Let us define f S 1 : M (X, θ) → S 1 the canonical extension of f S as a monoid homomorphism. That is, whenever w ∈ S (X, θ), f S 1 (w) = f S (w), and f S 1 (ǫ) = 1. Let w, w ′ ∈ S (X, θ) such that , and then we have Let S be a semigroup, and let f : X → S be a function that respects the commutations. If f is a R-invariant, then its canonical semigroup extension f S passes to the quotient S (X, θ) / * ⇐ ⇒ R .
Proof. Let w, w ′ ∈ S (X, θ) such that w * ⇐ ⇒ R w ′ . Then there are n > 0, w 0 , · · · , w n ∈ S (X, θ), w 0 = w, w n = w ′ such that for every 0 ≤ i < n, w i = w i+1 or w i ⇔ R w i+1 . Therefore for every 0 ≤ i < n, either We are now in position to establish the main result of this paper.
◮ Theorem 15. Let (X, θ) be a commutation alphabet, and let (S (X, θ), ⇒ R ) be an alphabetic rewriting system. There exist a group TG (X, θ, R) and a mapping t : X → TG (X, θ, R) that respects the commutations which is an R-invariant such that for every group G, and every (commutations respecting) R-invariant mapping f : X → G, there is a unique group homomorphism h : TG (X, θ, R) → G such that the following diagram commutes.
Proof. Let G be a group and let f : X → G be a commutations respecting R-invariant mapping. According to the universal problem of free partially commutative semigroups, because G is also a semigroup, we have the following commutative diagram.
According to corollary 14, we may complete the previous diagram in a natural way (the Now, let us assume that (S (X, θ), ⇒ R ) is convergent. Then, by proposition 11, S (X, θ) / * ⇐ ⇒ R is isomorphic to the free partially commutative semigroup S (Irr(X), θ Irr(X) ). Therefore, M = (S (X, θ) / * ⇐ ⇒ R ) 1 is isomorphic to the free partially commutative monoid M (X, θ) (by lemma 1). Finally, the Grothendieck group G(M ) is isomorphic to the Grothendieck group G (M (X, θ)) (because G(·) is functorial) so that it is isomorphic to the free partially commutative group G (X, θ) (by lemma 2). The fact that in this case, t is the normal form N • j S ,X : X → S (Irr (X), θ Irr (X) ) restricted to the alphabet X (where S (Irr (X), θ Irr (X) ) is naturally identified with a sub-semigroup of G (Irr (X), θ Irr (X) )) is quite obvious to check. ◭ ◮ Definition 16. The group TG (X, θ, R) is called the Tutte-Grothendieck group and t the universal Tutte-Grothendieck R-invariant of the alphabetic rewriting system (S (X, θ), ⇒ R ).

Some examples
This section is devoted to the presentation of several examples of Tutte-Grothendieck groups and universal invariants corresponding to convergent alphabetic rewriting systems. These examples come from the theory of graphs (Tutte polynomial), from algebra (Weyl algebra, and Poincaré-Birkhoff-Witt theorem) and from combinatorics (prefabs).

The Tutte polynomial
In its famous paper [19], Tutte used the following decomposition of (isomorphism classes of) finite multigraphs (graphs with multiple edges and loops). Let G be a multigraph, and e be a link (edge which is not a loop nor a bridge) in G. Let G − e be the graph obtained from G by erasing e, and let G/e be the graph obtained by contraction of e in G (e is removed, and its origin and source are identified). Then G is decomposed into (G − e) + G/e (+ being the free commutative juxtaposition). As explained in [8] in terms of an order relation, a rewriting system may be defined, and the universal invariant attached to this system is the well-known Tutte polynomials (see [19]).

Integral Weyl algebra
For any set X, let X ⊕ be the free commutative semigroup generated by X (that is, X ⊕ = S (X, θ, ), where θ = (X × X) \ ∆ X and ∆ X is the equality relation on X), written additively. Recall also that the free Abelian group generated by X, namely G (X, θ), is isomorphic to the group (under point-wise addition) Z (X) of all mappings from X to Z with a finite support (the support of a function f : X → Z is the set of all x ∈ X such that f (x) = 0), see for instance [6].
Moreover the alphabetic rewriting system (X ⊕ , ⇒ R ) is convergent (it is not difficult to check this property using for instance techniques from [3]). Let θ = (X × X) \ ∆ X . Then TG (X, θ, R) = G (Irr(X), θ Irr(X) ) = Z (Irr(X)) . Therefore we recover the well-known fact (see [13]) that the integral Weyl algebra A Z = Z a, b / I [a,b] with two generators (where Z a, b denotes the ring of the free monoid X = Y * = { a, b } * , and where I [a,b] is the two-sided ideal of Z a, b generated by ab − ba − 1) is free as an Abelian group with generators Irr(X). The universal Tutte-Grothendieck R-invariant t of (X ⊕ , ⇒ R ) is the normal form of the words in X = Y * . For instance, t(babab) = b 2 a 2 + 3b 2 a + b.
Let c be a variable (distinct from a, b) and Then we can check that (X ⊕ c , ⇒ Rc ) is a convergent alphabetic rewriting system whose Tutte-Grothendieck group is Z Irr(Xc) where Irr(X c ) = { c i b j a k : i, j, k ∈ N } (note that c i b j a k = c i1 b j c i2 a k c i3 for every non-negative integers i 1 , i 2 , i 3 , i = i 1 + i 2 + i 3 , j and k, since c commutes with all other elements). This gives us immediately a free Z-basis for the central extension Z a, b, c / I [a,b],c (where I [a,b],c is the two-sided ideal of the ring Z a, b, c of the monoid Y * c generated by ab − ba − 1 and cx − xc for every x ∈ { a, b }) of the integral Weyl algebra A Z .

The Poincaré-Birkhoff-Witt theorem
Let g be a Lie algebra over some basis ring 2 R which is free as an R-module (see [7]). Let B be a basis of g seen as a (free) R-module. Let us assume that B is linearly ordered by ≤. Let X = B * be the free monoid generated by B. Let R = { (uhgv, ughv) : g, h ∈ B, g < h, u, v ∈ B * } ⊆ X × X ⊕ . It is obvious that (X ⊕ , ⇒ R ) is a convergent alphabetic rewriting system. Moreover, Irr(X) = { g 1 · · · g n : n ≥ 0, g i ∈ B for all 0 ≤ i ≤ n, g i ≤ g i+1 for all 0 ≤ i < n } and its Tutte-Grotendieck group is Z (Irr(X)) , while its universal Tutte-Grothendieck invariant t is the re-ordering of an element of X in an increasing order (relative to ≤). We recognize the famous Poincaré-Birkhoff-Witt theorem ( [4,16,23]).

Prefabs
In [2], Bender and Goldman introduced the notion of a prefab, for combinatorial purposes (computation of some generating functions). We recall here (a part of) this concept. Let X be a set together with a multivalued binary operation • (meaning that x, y ∈ X implies that x • y ⊆ X) subjected to properties given below. For every x, y ∈ X, x • y is a finite set. The operation • is extended to the power set 2 X of X by A • B = { z ∈ S : x ∈ x • y for some x ∈ A, y ∈ B }. If x ∈ X and A ⊆ X, then we let x • A be equal to { x } • A = A • { x }, and x i is defined by induction: x 1 = { x }, and x i+1 = x • x i for every positive integer i. We say that (X, •) is a prefab if the composition • on 2 X is associative, commutative (therefore 2 X becomes a semigroup), and has an identity 3 i ∈ S such that x • i = { x } = i • x for every x ∈ X (then 2 X is a monoid). An element p ∈ X \ { i } is called a prime if p ∈ x • y implies x = i or y = i. We say that (X, •) is a unique factorization prefab if every x ∈ X \ { i } factors uniquely into primes in the sense that x ∈ p i1 1 • · · · • p in n for a unique set of n > 0 primes { p i : 1 ≤ i ≤ n } and unique positive integers i 1 , · · · , i n . We say that (X, •) is a very unique factorization prefab if x ∈ (p i1 1 • · · · • p im m ) • (q j1 1 • · · · • q jn n ) where m > 0, n > 0, all the i's and all the j's are positive integers, all the p's are mutually distinct primes, and all the q's are mutually distinct primes (but some q's may be equal to some p's), then there exist unique elements y ∈ p i1 1 • · · · • p im m and z ∈ q j1 1 • · · · • q jn n such that x ∈ y • z. In the original definition of a prefab, there is also a mapping f : 2 X → N which serves as a weigth function for a combinatorial use but which is not needed here.
Let (Y, •) be a unique and very unique factorization prefab. Let P be the set of primes of this prefab. Let X = Y \ { i }. Let R = { (x, y + z) : ∃y, z ∈ X, x ∈ y • z } ⊆ X × X ⊕ . According to the properties of unique factorization, very unique factorization, associativity and commutativity of •, it is clear that (X ⊕ , ⇒ R ) is a convergent alphabetic rewriting system. We have Irr(X) = P , and the Tutte-Grothendieck group is, as expected, Z (P ) . It is