A note on the monotonicity of reducible quantifiers (Q2268353)

Polyadic quantifiers are defined as functions mapping \(n\)-ary relations to truth values, thus \(Q\) is a type \(\langle n\rangle\) quantifier if \(R\) is an \(n\)-ary relation, \(R \in Q\) and \(Q(R)=1\). If \(Q\) is a type \(\langle n\rangle\) quantifier, then the Boolean complement \((\lnot Q)\), the post-complement \((Q\lnot)\), and the dual \(\lnot(Q\lnot)\) are defined. The composition of a type \(\langle k\rangle\) quantifier \(Q_1\) with a type \(\langle m\rangle\) quantifier \(Q_2\) is defined, as usual in the sense of \(\{0,1\}\)-valued \(n\)-ary relations, as being the type \(\langle k+m\rangle\) quantifier \(Q_1\circ Q_2(R) = Q_1(\{a_1, \dots a_k : Q_2(a_1, \dots a_kR)\})\), for any \((k+m)\)-ary relation \(R\). Reversely, the type \(\langle n\rangle\) quantifier \(Q\) is \((k,m)\)-reducible iff there exist a type \(\langle k\rangle\) quantifier \(Q_1\) and a type \(\langle m\rangle\) quantifier \(Q_2\) such that \(k+m=n\) and \(Q = Q_1\circ Q_2\). The results of this mathematical note provide necessary and sufficient conditions for the monotonicity and ``continuity'' of polyadic quantifiers to hold depending on the respective properties of their component quantifiers. The following definitions are necessary: Monotonicity: A type \(\langle n\rangle\) quantifier \(Q\) is monotone increasing (resp. decreasing) iff for any of its \(n\)-ary relations \(R \subseteq S\), \(R(S) = 1\) if \(Q(R) = 1\). Continuity: A type \(\langle n\rangle\) quantifier \(Q\) is continuous iff for any \(n\)-ary relations \(R, R_1, R_2\) such that \(R_1\subseteq R\subseteq R_2\), if \(Q(R_1) \land Q(R_2) = 1\) then \(Q(R) = 1\). Now, the main results of the paper can be given: Proposition 12 (p. 127): (i) The composed quantifier \(Q_1\circ Q_2\) is monotone increasing iff \(Q_1\) and \(Q_2\) are both monotone increasing or both monotone decreasing. (ii) The composed quantifier \(Q_1\circ Q_2\) is monotone decreasing iff one of \(Q_1\) or \(Q_2\) is monotone increasing and the other is monotone decreasing. (iii) \(Q_1\circ Q_2\) is not monotonic iff one of \(Q_1\) or \(Q_2\) is not monotonic. Proposition 13 (p. 128): If a reducible type \(\langle k+m\rangle\) quantifier \(Q_1\circ Q_2\) is continuous, where \(Q_1\) and \(Q_2\) are nontrivial, then \(Q_1\) and \(Q_2\) are both continuous. Notice that \(Q_i\) are nontrivial iff there exist relations \(R_i\neq\emptyset\) for which \(Q_i(R_i) = 1\), with \(i=1,2\).