\( stable(P) \wedge stable(Q) \quad \Rightarrow \quad stable(P \wedge Q) \)
All transitions from states in which \(P\) holds are to states in which \(P\) holds. All transitions from states in which \(Q\) holds are to states in which \(Q\) holds. Therefore all transitions from states in which \(P \wedge Q\) holds are to states in which \(P \wedge Q\) holds.
\( stable(P \wedge Q) \quad \Rightarrow \quad stable(P) \wedge stable(Q) \)
Suppose \(P\) holds in \(\{v, w\}\), and \(Q\) in \(\{w\}\), and therefore \(P \wedge Q\) holds in \(\{w\}\).
Assume transitions from \(w\) are to \(w\), and so \(stable(P \wedge Q)\).
The transition from \(v\) could be to a state other than \(v, w\). in which case \(\{v, w\}\) is not stable.
For example:
do x = 0 -> x = 0 x = 1 -> x = 2 od\(x = 0\), i.e., \(Q\) is stable but \((x = 0) \vee (x = 1)\), i.e. \(P\) is not stable. The left-hand side --- "\(stable(P \wedge Q)\) --- of the implication is true but the right-hand side --- \(stable(P) \wedge stable(Q)\) --- is false.
All states in all paths from states in which \((x = 0) \vee (x = 1)\) satisfy the invariant. Therefore all states in all paths from states in which \((x = 0)\) also satisfy the invariant.
K. Mani Chandy, Emeritus Simon Ramo Professor, California Institute of Technology