Not all \(p\)-vectors decompose (Q1567439)

A nonzero element of \(\Lambda ^{p}V\) (\(p\)-vector) \(u\) is said decomposable or pure if \ \(u=v_{1}\wedge \dots\wedge \) \(v_{p\text{ }}\)for some \( v_{1},\dots,v_{p}\in V\). Criteria for decomposability have been given by several authors, e.g. by \textit{W. Greub} [Multilinear algebra. 2nd ed. (1978; Zbl 0387.15001)], \textit{S. MacLane} and \textit{G. Birkhoff} [Algebra. (1967; Zbl 0153.32401)] and \textit{T. Yokonuma} [Tensor spaces and exterior algebra. (1992; Zbl 0754.15028)]. For an \(n\)-dimensional vector space \(V\) over a field \(K\), with \( n<\infty \), it is known that for \(p=1,n-1,n\) all \(\;p\)-vectors are decomposable. Using as key tool \ that there is an isomorphism \(\Lambda ^{n-p}V\rightarrow \Lambda ^{p}V\) , for \(1\leqslant p\leqslant n-1\), which sends the set of decomposable \((n-p)\)-vectors onto the set of decomposable \(p \)-vectors, the author establishes that if \(1\leqslant p\leqslant n-1\), then all nonzero \(p\)-vectors decompose if and only if \(p\) is either \(1\), \(n-1\) or \(n\). If \(V\) is an infinite-dimensional vector space then each \(p\)-vector decomposes if and only if \(p=1\). The result does not extend to a finitely generated commutative ring [cf. \textit{N. Bourbaki}, Elements of mathematics. Algebra I. Chapters 1-3. (1974; Zbl 0281.00006)]. Further if the field \(K\) is finite, \(K\cong \mathbb{F}_{q}\) , it is proved that then the number of decomposable vectors in \(\Lambda ^{p}V\) is \(d/e\), where \(d=(q^{n}-1)(q^{n}-q)(q^{n}-q^{2})\dots(q^{n}-q^{p-1})\) and \( e=(q^{p}-1)(q^{p}-q)(q^{p}-q^{2})\dots(q^{p}-q^{p-1})/(q-1)\) and the probability that a nonzero \(p\)-vector of \(\mathbb{F}_{q}^{(n)}\) decomposes is negligible.