Proof of Constants, Sums, Scalar Multiples and Differences of Differentiable Functions on an Open Subset of a Real Inner Product Space
lemmalem:c2-algebra-hilbert-2026aThe first-order and second-order expansions of a sum or a scalar multiple are obtained by splitting the remainder into the two remainders, or by factoring out the scalar; continuity of the resulting gradient and Hessian maps follows from the triangle inequality in the norm and in the form norm.
We use repeatedly that for , by Real Inner Product Space §distance and Elementary Identities in a Real Inner Product Space §homogeneity, and record two remarks.
Remark (i). If are continuous on as maps into and , then so are the maps and . Indeed, for the vector space identity and The Norm Metric of a Real Inner Product Space: Triangle Inequalities, Limits and Continuity §triangle give
so that, given a positive , radii for and for at the positive number (claim 8 of Elementary Order Arithmetic in an Ordered Field) and the lesser of the two (claim 9 there) witness continuity of the sum at relative to , by Continuous Map Between Metric Spaces. For the scalar multiple, by The Norm Metric of a Real Inner Product Space: Triangle Inequalities, Limits and Continuity §metric; if the map is constant, and otherwise a radius for at serves.
Remark (ii). The same holds for maps continuous into : by Elementary Properties of Bounded Symmetric Bilinear Forms: Norm, Quadratic Form, Order and Continuity §vector-space the set is a vector space, so and , and Elementary Properties of Bounded Symmetric Bilinear Forms: Norm, Quadratic Form, Order and Continuity §norm-axioms gives
The argument of remark (i) then applies verbatim.
Claim 1. Let and let have constant value , and let . Since is open, Open Subset of a Metric Space provides a positive with , so that whenever . For such , by Elementary Identities in a Real Inner Product Space §zero, and for every positive by claim 5 of Elementary Arithmetic in an Ordered Field; taking for every shows that is differentiable at with gradient . The gradient map is constant, hence continuous on , so by The Classes and on an Open Subset of a Real Inner Product Space §c1. Moreover is differentiable at every point of and, for all with ,
so taking for every shows that is a second derivative of at . The Hessian map is constant, hence continuous, and by The Classes and on an Open Subset of a Real Inner Product Space §c2.
Claim 2. Suppose and are differentiable at , with gradients and . Let be positive, let and be radii supplied by Fréchet Differentiability and the Gradient on an Open Subset of a Real Inner Product Space §differentiable for the positive number , and let be the lesser of the two, which is positive. Every with satisfies and, using (condition (b) of Real Inner Product Space §inner-product), claim 5 of Properties of the Absolute Value in an Ordered Field and claim 8 of Elementary Order Arithmetic in an Ordered Field,
Hence is differentiable at with gradient .
Suppose in addition that and have second derivatives and at , with radii and as in The Second Derivative and the Hessian on an Open Subset of a Real Inner Product Space §expansion, and let be the lesser of the two, which is positive. Both and are differentiable at every point of , hence so is by the previous paragraph, with for every such . The form lies in by Elementary Properties of Bounded Symmetric Bilinear Forms: Norm, Quadratic Form, Order and Continuity §vector-space. Given a positive , let be the lesser of and of the two radii supplied by that clause for ; for all with , the identity
which uses condition (b) of Real Inner Product Space §inner-product and Bounded Symmetric Bilinear Forms on a Real Inner Product Space: Norm, Order, Identity Form and Restriction §identity, together with claim 5 of Properties of the Absolute Value in an Ordered Field, bounds the left-hand side in absolute value by . Hence is a second derivative of at .
Finally, if then is differentiable on with gradient map , continuous by remark (i), so ; and if then in addition has at every the second derivative , and its Hessian map is continuous by remark (ii), so .
Claim 3. If then is the function with constant value , and claim 1 applies; moreover and by claim 3 of Elementary Identities in a Vector Space, applied in and in the vector space of Elementary Properties of Bounded Symmetric Bilinear Forms: Norm, Quadratic Form, Order and Continuity §vector-space, so the asserted formulas hold.
Suppose , so that by claim 1 of Properties of the Absolute Value in an Ordered Field, and let be differentiable at with gradient . Given a positive , let be the radius supplied by Fréchet Differentiability and the Gradient on an Open Subset of a Real Inner Product Space §differentiable for the positive number . For with , condition (c) of Real Inner Product Space §inner-product gives , so claim 4 of Properties of the Absolute Value in an Ordered Field and claim 5 of Elementary Arithmetic in an Ordered Field give
Hence is differentiable at with gradient . The second-derivative assertion follows in the same way from the identity
where and by Elementary Properties of Bounded Symmetric Bilinear Forms: Norm, Quadratic Form, Order and Continuity §vector-space, using condition (c) of Real Inner Product Space §inner-product and Bounded Symmetric Bilinear Forms on a Real Inner Product Space: Norm, Order, Identity Form and Restriction §identity. The class assertions follow from remarks (i) and (ii) as in claim 2.
Claim 4. For every one has by claim 5 of Elementary Identities in a Vector Space applied in the vector space , so is the function . Claims 2 and 3 therefore give every assertion, the gradient being and the second derivative , by claim 5 of Elementary Identities in a Vector Space in and by Bounded Symmetric Bilinear Forms on a Real Inner Product Space: Norm, Order, Identity Form and Restriction §identity in .
Loading…
Prerequisites
008521ea-77d5-4be3-a35c-caaf4df5a674