Roy S. Freedman
Abstract:Given a set of N propositions, if any pair is mutual exclusive, then the set of all propositions are N-way jointly mutually exclusive. This paper provides a new general counterexample to the converse. We prove that for any set of N propositional variables, there exist N propositions such that their N-way conjunction is zero, yet all k-way component conjunctions are non-zero. The consequence is that N-way joint mutual exclusion does not imply any pairwise mutual exclusion. A similar result is true for sets since propositional calculus and set theory are models for two-element Boolean algebra.
What problem does this paper attempt to address?
The core problem that this paper attempts to solve is to explore whether the N - way joint mutual exclusion of N propositions implies pairwise mutual exclusion between any two propositions. Specifically, the author attempts to prove that the N - way joint mutual exclusion of N propositions does not mean that any two of these propositions are pairwise mutually exclusive.
### Background and Problem Description
In logic and set theory, two propositions \(X_1\) and \(X_2\) are mutually exclusive if they cannot be true simultaneously, that is:
\[ X_1 \land X_2 = 0 \]
For three propositions \(X_1, X_2, X_3\), if any pair is mutually exclusive, then the triple conjunction of these three propositions (i.e., the situation where all three are true simultaneously) is also zero:
\[ X_1 \land X_2 \land X_3 = 0 \]
More generally, for N propositions, if any pair is mutually exclusive, then the N - fold conjunction of these N propositions is also zero. However, does the converse hold? That is, does the N - fold conjunction of N propositions being zero imply that any pair of propositions is mutually exclusive?
### Main Contributions of the Paper
The paper proves that N - way joint mutual exclusion does not imply pairwise mutual exclusion by constructing a counterexample. Specifically, the author proves that for any N propositional variables \(A_1, A_2,\ldots, A_N\), there exist N non - zero propositions \(P_1, P_2,\ldots, P_N\) such that:
- The N - fold conjunction of N propositions is zero:
\[ P_1 \land P_2 \land \cdots \land P_N = 0 \]
- But the k - fold conjunction of any k propositions is not zero (where \(2\leq k < N\)):
\[ P_{i_1} \land P_{i_2} \land \cdots \land P_{i_k} \neq 0 \]
### Construction of the Counterexample
The author achieves this by constructing specific propositions \(P_i\). Each proposition \(P_i\) is the disjunction of a set of "exclusion terms", each exclusion term contains \(N - 1\) propositional variables, and exactly one variable is negated. For example, in the case of \(N = 4\), the four propositions are respectively:
\[
\begin{aligned}
P_1 &= A_2' A_3 A_4+A_2 A_3' A_4+A_2 A_3 A_4'\\
P_2 &= A_1' A_3 A_4+A_1 A_3' A_4+A_1 A_3 A_4'\\
P_3 &= A_1' A_2 A_4+A_1 A_2' A_4+A_1 A_2 A_4'\\
P_4 &= A_1' A_2 A_3+A_1 A_2' A_3+A_1 A_2 A_3'
\end{aligned}
\]
These propositions satisfy the above properties, that is, the N - fold conjunction of N propositions is zero, but the conjunction of any two propositions is not zero.
### Conclusion
The paper proves through this counterexample that N - way joint mutual exclusion does not imply pairwise mutual exclusion, thus solving this important problem in logic and set theory. This result is not only meaningful in propositional logic, but also has similar applications in probability theory and set theory.