Existence of a grothendieck group

1. Proposition

Let \(M\) be a commutative monoid, then there exists a grothendieck group \(K\)

2. Proof

Let \(M \times M\) with the equivalence relation

\begin{align*} (m_1, m_2) \sim (n_1,n_2) :\Leftrightarrow \exists t \in M : m_1 + n_2 + t = n_1 + m_2 + t \end{align*}

2.1. reflexive

follows any \(t\), e.g. \(t = 0\) and

\begin{align*} m_1 + n_2 + 0 = m_1 + n_2 + 0 \end{align*}

2.2. symmetric

follows from the symmetry of \(=\)

2.3. transitive

Let \((m_1, m_2) \sim (n_1,n_2) \sim (l_1,l_2)\) with

\begin{align*} m_1 + n_2 + t =& n_1 + m_2 + t \\ n_1 + l_2 + s =& l_2 + n_2 + s \end{align*}

for suitable \(l,s \in M\) Thus

\begin{align*} m_1 + n_2 + t + n_1 + l_2 + s =& n_1 + m_2 + t + l_1 + n_2 + s \\ m_1 + l_2 + (n_1 + n_2 + t + s) =& l_1 + m_2 + (n_1 + n_2 + t + s) \end{align*}

and hence

\begin{align*} (m_1,m_2) \sim (l_1,l_2) \end{align*}

2.4. welldefined binary operation

Let

\begin{align*} +: G(M) \times G(M) \rightarrow& G(M) \\ ([x,y], [x',y']) \mapsto& [x + x', y + y'] \end{align*}

Suppose \([x_1,x_2] \sim [x_1',x_2']\) and \([y_1,y_2] \sim [y_1',y_2']\) Then there exist \(s,t \in M\) such that

\begin{align*} x_1 + x_2' + t =& x_1' + x_2 + t \\ y_1 + y_2' + s =& y_1' + y_2 + s \end{align*}

Therefore,

\begin{align*} (x_1 + y_1) + (x_2' + y_2') + (s + t) =& (x_1' + y_1') + (x_2 + y_2) + (s + t) \end{align*}

or equivalently

\begin{align*} [x_1 + y_1, x_2 + y_2] =& [x_1' + y_1', x_2' + y_2'] \\ [x_1,x_2] + [y_1,y_2] =& [x_1',x_2'] + [y_1',y_2'] \end{align*}

2.5. commutativity

follows from

\begin{align*} [x,y] + [x',y'] =& [x + x', y + y'] \\ =& [x' + x, y' + y] \\ =& [ \end{align*}

2.6. neutral element

follows from

\begin{align*} [e,e] + [x,y] =& [x + e, y + e] \\ =& [x, y] \end{align*}

and the commutativity

2.7. inverse element

for \([x,y]\), the inverse element is \([y,x]\), as

\begin{align*} [x,y] + [y,x] =& [x + y, y + x]\\ =& [x + y, x + y] \end{align*}

and furthermore

\begin{align*} (x + y) + 0 + 0 =& 0 + (x + y) + 0 \end{align*}

thus

\begin{align*} [x + y,x + y] = [e,e] \end{align*}

2.8. universal property

Let \(\varphi: M \rightarrow H\) be a monoid-homomorphism for an abelian group \(H\). Then let

\begin{align*} \psi: G(M) \rightarrow& H \\ [x,y] \mapsto& \varphi(x) - \varphi(y) \end{align*}

2.8.1. commutes

For the canonical inclusion

\begin{align*} \iota: M \rightarrow& G(M) \\ m \mapsto& [m,0] \end{align*}

it follows, that

\begin{align*} \varphi(m) = \varphi(m) - 0 \\ =& \varphi(m) - \varphi(0) \\ =& \psi([m,0]) \\ =& \psi(\iota(m)) \end{align*}

2.8.2. unique

Let \(\psi': G(M) \rightarrow H\) be a group homomorphism. Suppose \(\psi'\) makes the diagram commute, then by assumption

\begin{align*} \psi'(\iota(m)) = \psi'([m,0] = \varphi(m) \end{align*}

Furthermore, since \([m,0] = -[0,m]\), it follows, that since \(\psi'\) is a group homomorphism

\begin{align*} \psi'[0,m] = - \varphi(m) \end{align*}

Hence \(\psi\) is unique, as

\begin{align*} \psi'[m,n] =& \psi'[m,0] + \psi[0,n] \\ =& \varphi(m) - \psi[n,0] \\ =& \varphi(m) - \varphi(n) \end{align*}

Date: nil

Author: Anton Zakrewski

Created: 2024-10-13 So 18:31