Theorem 11.3.2. (The convex representative theorem). A function has positive distributional Hessian if and only if its almost-everywhere class contains a finite convex function . Such a representative is unique. Every standard compactly supported nonnegative smooth approximate identity gives
Argument in the manuscript
Proof. Assume first that the Hessian is positive. On the inner domain where convolution is defined, is smooth and its Hessian is positive semidefinite: for each vector , its quadratic Hessian coefficient is the convolution of the nonnegative distribution with a nonnegative kernel. Thus is convex on every convex inner subdomain.
The convolutions converge to in on every relatively compact open subset. Their norms are therefore uniformly bounded there for sufficiently small . Cover a given compact subset by finitely many balls whose fourfold enlargements remain in . Lemma 11.3.1 gives uniform bounds and equicontinuity on the compact subset. Arzelà–Ascoli and a diagonal argument over an exhaustion give a subsequence converging locally uniformly to a finite continuous function . The convex inequality passes to the limit on each compact line segment, so is convex. Its limit is also , hence almost everywhere.
Two finite convex representatives are continuous. If they differed at one point, they would differ on a nonempty open set, which has positive Lebesgue measure. Hence they coincide. This uniqueness and the same local compactness argument show convergence of the whole family , not just the selected subsequence: any sequence violating uniform convergence on a compact set would have a locally uniformly convergent subsequence with the same unique limit.
Conversely, if almost everywhere for a finite convex , its convolutions are smooth convex functions and converge locally uniformly to . Their positive Hessian distributions converge to the Hessian distribution of . Positivity is closed under convergence on tests, proving the converse. ◻