On the generation of clones containing near-unanimity operations (Q535111)

For \(d \geq 2\), a \((d+1)\)-ary operation \(m\) on a set \(A\) is a near-unanimity operation if \[ m(x,y,y,\dots,y) = m(y,x,y,\dots ,y) = m(y,y,\dots ,y,x) = y \] for all \(x,y \in A\). The well-known Baker-Pixley Theorem [\textit{K. A. Baker} and \textit{A. F. Pixley}, Math. Z. 143, 165--174 (1975; Zbl 0292.08004)] says that if \(C\) is a clone on \(A\) containing a \((d+1)\)-ary near-unanimity operation then every operation that preserves the set of all \(d\)-ary relations on \(A\) that are preserved by \(C\), is in \(C\). As an easy consequence, such clones are finitely generated and only finitely many clones can contain near-unanimity operations of a given arity. This leads to the question concerning arity of the generating functions. Some results concerning this question were provided by \textit{H. Lakser} [Algebra Univers. 26, 78--89 (1989; Zbl 0671.08002)]. In the present paper, the author shows that, for arbitrary \(d \geq 2\) and large enough \(n\), the number \((n-1)^d - 1\) is the smallest integer \(k\) such that, for every clone \(C\) on an \(n\)-element set that contains a \((d+1)\)-ary near-unanimity operation, \(C^{(k)}\) generates \(C\).