Open Mapping and Closed Graph Theorems: A Simple Intro

The open mapping theorem says a surjective bounded linear operator between Banach spaces maps open sets to open sets. The closed graph theorem says a linear operator between Banach spaces whose graph is closed must be bounded. Together with the uniform boundedness principle they show that completeness plus linearity forces strong regularity.

They matter because they replace hard analytic checks with easy ones. The bounded inverse theorem (a corollary) gives a continuous inverse for free once an operator is bijective and bounded. The closed graph theorem lets you check the easy sequential condition, x_n -> x and T(x_n) -> y imply y = T(x), instead of proving boundedness directly.

Key Concepts

• Open map images of open sets are open; surjective bounded linear maps between Banach spaces qualify.
• Bounded inverse theorem: a bijective bounded linear operator between Banach spaces has a bounded inverse; the most-used corollary.
• Graph of T: the set {(x, T(x))}; T has closed graph if x_n -> x and T(x_n) -> y force y = T(x).
• Closed graph theorem: a linear map between Banach spaces with closed graph is bounded.
• Completeness is essential both theorems fail without completeness on both spaces; Baire category drives the proofs.

How It Works

The open mapping proof uses Baire category: surjectivity writes Y as a countable union of closures of T(B(0, n)), so one has nonempty interior, and linearity promotes this to T(B(0, 1)) containing a ball around 0, which scales to openness everywhere. The bounded inverse theorem follows since a bijective open map has a continuous inverse. For the closed graph theorem, the graph norm ||x||_T = ||x|| + ||T(x)|| makes X a Banach space, and the identity map to X is bijective and bounded, so the bounded inverse theorem makes the norms equivalent, bounding T. In practice you apply it to differential operators, where the graph condition needs only pointwise convergence.

Worked Example

See the bounded inverse theorem on T(x) = 3x from R to R. T is linear, bounded with ||T|| = 3, and bijective. The open mapping theorem says T is open: indeed T maps B(0, r) = (-r, r) to (-3r, 3r) = B(0, 3r). The inverse S(y) = y/3 exists algebraically, and the theorem guarantees it is bounded with no extra estimates; directly, ||S|| = 1/3. Contrast incomplete spaces: on polynomials with the sup norm, differentiation is bijective with closed graph yet unbounded, so completeness cannot be dropped.

Common Misconceptions

• "Open map means the same as continuous." Continuity is preimages of open sets being open; an open map sends open sets to open sets. They are independent: some continuous maps are not open, some open maps are not continuous.
• "Closed graph means the graph is a closed set, which is automatic." For discontinuous linear maps the graph need not be closed; closedness is a genuine restriction the theorem converts into boundedness.
• "The theorems hold for any normed spaces." Completeness of both domain and codomain is essential, as the differentiation counterexample on incomplete polynomial spaces shows.

Real-World Applications

• Differential equations solution operators for linear PDEs are shown bounded via the closed graph theorem.
• Numerical analysis: stability of a discretization scheme often follows from a uniform boundedness argument on Banach spaces.
• Signal processing: the bounded inverse theorem guarantees an invertible filter has a stable (bounded) inverse filter.
• Quantum mechanics Hellinger-Toeplitz: an everywhere-defined symmetric operator on Hilbert space is bounded.

Open mapping gives bounded inverses for free; closed graph turns a sequential check into boundedness. Both need Banach spaces and replace hard estimates with soft arguments.