TheoremBase

Proof sketch: by induction on the length, a dependent spanning family has a vector in the span of the preceding ones, which can be dropped without changing the span.

Proof

Routine facts about equality of tuples are used without comment (Toolkit for AL2). We prove the claim by induction on n∈N0n\in\mathbb{N}_{0} (Induction on β„•β‚€ from 0 for a Condition Expressed by a Predicative Formula (ii)), for all spanning nn-tuples at once: the statement proved for nn is that (i) holds for every nn-tuple in VV that spans VV. In both the base case and the induction step, a spanning family vv that is linearly independent is itself a basis (Finite Basis of a Vector Space (i)), and we may take m=nm=n and k=id[n]k=\mathrm{id}_{[n]}.

For n=0n=0 this always applies: the empty family is linearly independent by Elementary Properties of Linear Independence (vi), so, as it spans VV, it is a basis, and we take m=0m=0 and kk the empty map.

For the induction step, assume the claim for nn, and let v∈Vn+1v\in V^{n+1} span VV. By the remark above we may assume that vv is linearly dependent. Then Elementary Properties of Linear Independence (iv) gives j∈[n+1]j\in[n+1] with vj∈span⁑(v1,…,vjβˆ’1)v_{j}\in\operatorname{span}(v_{1},\dots,v_{j-1}). Let v(j)v^{(j)} be the nn-tuple of the other vectors of vv, as in Dropping a Redundant Vector Does Not Change the Span. For i<ji<j we have gj(i)=ig_{j}(i)=i (Bijections between Intervals: the Gap Map and the Reversal (i)), so vi=vi(j)v_{i}=v^{(j)}_{i}; hence span⁑(v1,…,vjβˆ’1)βŠ†span⁑(v(j))\operatorname{span}(v_{1},\dots,v_{j-1})\subseteq\operatorname{span}(v^{(j)}) by The Span of a Finite Family is the Smallest Subspace Containing It (ii) and The Span of a Finite Family is the Smallest Subspace Containing It (iii). Therefore vj∈span⁑(v(j))v_{j}\in\operatorname{span}(v^{(j)}), and by Dropping a Redundant Vector Does Not Change the Span (i) the nn-tuple v(j)v^{(j)} spans VV.

The induction hypothesis, applied to v(j)v^{(j)}, gives m≀nm\leq n and a map kβ€²:[m]β†’[n]k':[m]\to[n], strictly increasing in the sense of (i), such that (vkβ€²(1)(j),…,vkβ€²(m)(j))(v^{(j)}_{k'(1)},\dots,v^{(j)}_{k'(m)}) is a basis of VV. Put k=gj∘kβ€²:[m]β†’[n+1]k=g_{j}\circ k':[m]\to[n+1]. It is strictly increasing because gjg_{j} is, by its definition in Bijections between Intervals: the Gap Map and the Reversal (i), and vk(l)=vkβ€²(l)(j)v_{k(l)}=v^{(j)}_{k'(l)} for every l∈[m]l\in[m], so (vk(1),…,vk(m))(v_{k(1)},\dots,v_{k(m)}) is that same basis; and m≀n+1m\leq n+1. This completes the induction and proves (i).

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…