next up previous
Next: aa_subdirect Up: aa_subdirect Previous: aa_subdirect

5. The proof of the subdirect representation theorem

5.1 Lemma. Given $ a \neq b$ in $ {\cal A}$, there exists a congruence relation $ \theta$ maximal with respect to the property $ a \not
\equiv b \; (\theta)$.

Proof. Let $ {\cal S} = \{\theta \in$   Con$ ({\cal A}) : \langle a,b
\rangle \not \in \theta\}$. Then $ {\cal S}$ is not empty, since $ 0 \in
{\cal S}$. Suppose $ {\cal C}$ is a chain of members of $ {\cal S}$, where each relation is regarded as a subset of $ {\cal A} \times {\cal A}$. Then $ \bigcup _ {\theta \in {\cal C}} \theta \in {\cal S}$, since all aspects of being in $ {\cal S}$ (specifically, being an equivalence relation, being compatible with the operations of $ {\cal A}$, and not containing $ \langle a,b \rangle$) can be checked using finitely many elements at a time and so can be checked inside just one member of $ {\cal C}$ at a time. Then by Zorn's Lemma, $ {\cal S}$ has a maximal member. $ \Box$



Let $ \theta _ {ab}$ be one such congruence relation maximal with respect to not identifying $ a$ and $ b$. Here $ \theta _ {ab}$ is in contrast to con$ (a,b)$, the smallest congruence relation that identifies $ a$ and $ b$. In fact, $ \theta _ {ab}$ can be described as a $ \theta$ maximal with respect to the property $ \theta \not \geq$   con$ (a,b)$.



5.2 Observation. For $ a \neq b$ in $ {\cal A}$, in Con$ ({\cal A})$ there is a least element $ > \theta _ {ab}$, namely $ \theta _ {ab} \vee$   con$ (a,b)$.



5.3 Observation. $ {\cal A}/\theta _ {ab}$ is subdirectly irreducible. Indeed, by Observation 1 and the Correspondence Theorem, Con$ ({\cal A}/\theta _ {ab})$ has a least element $ > 0$ and so is subdirectly irreducible.



5.4 Observation. $ \bigcap _ {a \neq b} \theta _
{ab} = 0$ in Con$ ({\cal A})$, where $ a,b$ range over $ {\cal A}$.



Proof of the Representation Theorem. By Observation [*] we have $ {\cal A} \hookrightarrow \prod _ {a \neq b} {\cal A}/\theta _ {ab}$, and by Observation [*] each $ {\cal A}/\theta _ {ab}$ is subdirectly irreducible.




next up previous
Next: aa_subdirect Up: aa_subdirect Previous: aa_subdirect
Kirby A. Baker 2003-02-21