TheoremBase

Sketch: the line is the enlarged subspace of the one-step extension lemma for the zero subspace and the zero functional; linearity of t x₀ ↦ t c is distributivity, and domination follows from nonnegative homogeneity of p applied to x₀ or to -x₀.

Proof

First, Rx0\mathbb{R}x_{0} is a linear subspace. It contains 0X=0 x00_{X}=0\,x_{0} by Elementary Identities in a Vector Space (ii), and for t,s,λ∈Rt,s,\lambda\in\mathbb{R},

t x0+s x0=(t+s) x0,λ(t x0)=(λt) x0t\,x_{0}+s\,x_{0}=(t+s)\,x_{0},\qquad \lambda(t\,x_{0})=(\lambda t)\,x_{0}

by Vector Space over a Field (viii) and Vector Space over a Field (v); so Rx0\mathbb{R}x_{0} satisfies Linear Subspace of a Real Vector Space (i), Linear Subspace of a Real Vector Space (ii) and Linear Subspace of a Real Vector Space (iii). The coefficient of a point of Rx0\mathbb{R}x_{0} is unique: if t x0=t′ x0t\,x_{0}=t'\,x_{0}, then

(t−t′) x0=t x0+(−t′) x0=t x0−t′ x0=t x0−t x0=0X(t-t')\,x_{0}=t\,x_{0}+(-t')\,x_{0}=t\,x_{0}-t'\,x_{0}=t\,x_{0}-t\,x_{0}=0_{X}

by Vector Space over a Field (viii), Elementary Identities in a Vector Space (iv) with the definition of differences, and Elementary Identities in a Vector Space (vii). Since x0≠0Xx_{0}\neq0_{X}, Elementary Identities in a Vector Space (xii) gives t−t′=0t-t'=0, that is, t=t′t=t'. This proves (i).

By (i), each point of Rx0\mathbb{R}x_{0} is t x0t\,x_{0} for exactly one tt, so gg is well defined. It is linear: by Vector Space over a Field (viii) and Vector Space over a Field (v),

g(t x0+s x0)=g((t+s) x0)=(t+s) c,g(λ(t x0))=g((λt) x0)=λt c,g(t\,x_{0}+s\,x_{0})=g\bigl((t+s)\,x_{0}\bigr)=(t+s)\,c,\qquad g\bigl(\lambda(t\,x_{0})\bigr)=g\bigl((\lambda t)\,x_{0}\bigr)=\lambda t\,c,

which are Linear Functional on a Linear Subspace (i) and Linear Functional on a Linear Subspace (ii).

Finally, gg is dominated by pp. Let t∈Rt\in\mathbb{R}. If t≥0t\ge0, the hypothesis c≤p(x0)c\le p(x_{0}) and Sublinear Functional (ii) give

g(t x0)=t c≤t p(x0)=p(t x0).g(t\,x_{0})=t\,c\le t\,p(x_{0})=p(t\,x_{0}).

If t<0t<0, put s=−t>0s=-t>0. Then s(−x0)=(−s) x0=t x0s(-x_{0})=(-s)\,x_{0}=t\,x_{0} by Elementary Identities in a Vector Space (iv), and the hypothesis −c≤p(−x0)-c\le p(-x_{0}) together with Sublinear Functional (ii) gives

g(t x0)=s (−c)≤s p(−x0)=p(s(−x0))=p(t x0).g(t\,x_{0})=s\,(-c)\le s\,p(-x_{0})=p\bigl(s(-x_{0})\bigr)=p(t\,x_{0}).

This proves (ii).

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…