Open map - Ciallo Wiki

1 Basic definitions and formulations

1.1 Open maps in topological spaces

Let \(X\) and \(Y\) be topological spaces and \(f: X \to Y\) a map. The map \(f\) is called open if for every open set \(U \subseteq X\), the image \(f(U)\subseteq Y\) is open. This definition does not require \(f\) to be continuous, though openness is often studied together with continuity because both properties control how sets behave under the map.

If \(f\) is open and surjective, then \(Y\) is covered by open sets of the form \(f(U)\), with \(U\) open in \(X\). When \(f\) is not surjective, the definition still requires that images of open sets be open in the ambient space \(Y\).

1.2 Equivalent characterizations

Several reformulations are frequently used.

Basis criterion. If \(\mathcal{B}\) is a basis for \(X\), then \(f\) is open if and only if \(f(B)\) is open in \(Y\) for every \(B\in \mathcal{B}\).

This reduces verification to smaller families of sets.

Neighborhood form. For each \(x\in X\) and each open neighborhood \(U\) of \(x\), the set \(f(U)\) is an open neighborhood of \(f(x)\) (whenever \(f(x)\in f(U)\), which holds for any neighborhood \(U\) of \(x\)).

In practice this means that open sets around \(x\) generate open “thickening” around \(f(x)\).

Image-of-interiors inclusion. For a map \(f\), openness implies that

\[ f(\operatorname{int}_X A)\subseteq \operatorname{int}_Y f(A) \] for all \(A\subseteq X\), where \(\operatorname{int}_X\) denotes interior in \(X\). This is a useful consequence, though the inclusion may fail without openness.

1.3 Relationship to continuity and embeddings

Openness is logically independent from continuity.

A map can be continuous but not open (e.g., an inclusion of a subspace with the induced topology typically fails to be open unless the subspace is open in the ambient space).
A map can be open without being continuous (though such examples often require carefully chosen topologies).

If \(f\) is a homeomorphism—a bijection that is continuous with continuous inverse—then \(f\) is both open and closed. More generally, when \(f\) is injective and continuous, openness ensures that \(f\) behaves like an embedding onto an open image; if the image is open, the map can resemble topological inclusion in how it preserves local openness.

1.4 Open-map examples in familiar spaces

Open projections from products. The projection \(\pi_Y: X\times Y \to Y\) is open when \(X\times Y\) carries the product topology. Indeed, any basic open set \(U\times V\) projects to \(V\), which is open.

Linear surjections between finite-dimensional spaces. In \(\mathbb{R}^n\), a surjective linear map is open: it maps open balls to sets containing open balls in the codomain (up to a suitable radius determined by the operator). This intuition extends to affine surjections.

Quotient maps. The canonical quotient map \(q: X \to X/{\sim}\) is open when the quotient topology is used, because images of open sets are defined to be open in the quotient space.

Coordinatewise maps. Maps of the form \((x_1,\dots,x_n)\mapsto (g_1(x_1),\dots,g_n(x_n))\) can be open under mild conditions on each coordinate map and the product topology, particularly when each \(g_i\) is open on its factor.

2 Connections with other map properties

2.1 Closed maps and duality considerations

A map \(f: X\to Y\) is closed if it sends closed sets in \(X\) to closed sets in \(Y\). Openness and closedness are dual notions via complements, but they are not equivalent in general unless additional assumptions are present (such as bijectivity with a continuous inverse).

For bijective continuous maps, failure of openness can correspond to failure of closedness, and vice versa. A common theme is that stronger structural hypotheses—especially homeomorphisms—force both properties.

2.2 Homeomorphisms and open/closed criteria

If \(f: X\to Y\) is a homeomorphism, then \(f\) is open and closed simultaneously. Conversely, if \(f\) is bijective and continuous and either open or closed, then \(f^{-1}\) is continuous, making \(f\) a homeomorphism. This criterion is a standard tool for identifying when a map is topologically “exact” rather than merely set-theoretically bijective.

2.3 Quotient maps and openness of induced maps

Given a surjection \(q: X\to Z\) that is a quotient map, a function \(f: Z\to Y\) is continuous precisely when \(f\circ q\) is continuous. Openness interacts with this picture as follows:

If \(f\circ q\) is open and \(q\) is surjective with appropriate quotient topology behavior, then \(f\) can often be shown open by translating openness back through \(q\).
In many constructions, checking openness of induced maps reduces to understanding images of saturated open sets under \(q\).

Although the details depend on the specific quotient structure, the quotient framework frequently provides canonical open maps and simplifies the study of maps defined on equivalence classes.

2.4 Local homeomorphisms and covering-type behavior

A local homeomorphism is a map \(f: X\to Y\) such that each point \(x\in X\) has a neighborhood \(U\) with \(f	_U: U\to f(U)\) a homeomorphism. Local homeomorphisms are automatically open because open sets can be pulled apart into pieces on which \(f\) acts homeomorphically, and homeomorphisms preserve openness.

Covering maps (in settings like topological manifolds) are a typical source of local homeomorphisms and thus exhibit strong openness properties.

3 Open mapping in analysis

3.1 Open mapping behavior for linear operators

In analysis, one studies maps between normed spaces where the topology comes from a norm. A linear operator \(T: X\to Y\) that is surjective is often expected to be open, but the statement requires hypotheses: in incomplete spaces, surjectivity alone may not guarantee openness.

When the relevant spaces are Banach (complete normed spaces), standard open mapping theorems yield powerful conclusions: surjective continuous linear maps become open. This connects the geometric structure of images (open sets) to the analytic structure (boundedness and completeness).

3.2 Normed spaces and topological vector spaces

For normed spaces, openness is compatible with the structure of the metric topology. In topological vector spaces, one typically works with translation invariance: if \(f\) is linear, then the openness of \(f\) can be tested at neighborhoods of \(0\). The key point is that many neighborhoods are “balanced” and defined by norms, making geometric arguments possible.

If a map respects the vector-space structure (e.g., linear or affine), openness often translates into estimates controlling how large an image of a neighborhood must be in the codomain.

3.3 Consequences for images and interior points

Openness has immediate geometric consequences.

If \(f\) is open, then for any open \(U\subseteq X\), the set \(f(U)\) contains an open neighborhood around each point of \(f(U)\) that comes from images of points in \(U\).
For many applications, this yields interior information: if \(U\) is open and \(f\) is open, then \(f(U)\) is not merely large—it is topologically thick (open).

In functional analysis, these interior properties underpin existence arguments for solutions to operator equations, showing that “approximate” solvability can imply actual solutions lying in robust regions rather than isolated points.

3.4 Where openness fails: common counterexamples

Openness can fail even when maps look “reasonable.”

Non-surjective maps. A continuous linear injection can map open sets to sets that are not open in the ambient codomain because the image may lie in a lower-dimensional subspace with empty interior.
Incomplete spaces without completeness hypotheses. Some surjective bounded linear operators between incomplete normed spaces are not guaranteed to be open; completeness is often essential in the classic theorems.
Nonlinear pathologies. For nonlinear maps, openness may fail due to folding, collapsing, or mapping regions into “thin” subsets (e.g., images with empty interior), even if the map is continuous.

These counterexamples illustrate that openness is a strong global property, not a mere local feature.

4 Open mapping theorems (statement-level overview)

4.1 The Banach–Schauder open mapping theorem

A central result in functional analysis is:

If \(X\) and \(Y\) are Banach spaces and \(T: X\to Y\) is a bounded linear operator that is surjective, then \(T\) is an open map.

Equivalently, \(T\) maps open subsets of \(X\) to open subsets of \(Y\).

This theorem converts the analytic assumptions (boundedness and completeness) into strong topological information about images.

4.2 Connections to the bounded inverse theorem

The open mapping theorem is tightly related to the bounded inverse theorem:

If \(T: X\to Y\) is a bounded linear bijection between Banach spaces, then \(T^{-1}\) is bounded (and thus continuous).
Openness of \(T\) and continuity of \(T^{-1}\) are intertwined: when \(T\) is bijective and continuous, proving openness (or closedness) helps establish that the inverse respects the topology quantitatively.

At a conceptual level, theorems form a triangle: open mapping ↔ bounded inverse ↔ continuity properties strengthened by completeness.

4.3 Applications to surjectivity and existence of solutions

Open mapping theorems are often used to establish that solutions to operator equations exist in a stable way.

If one considers an equation \(T x = y\) with \(T\) bounded linear and surjective, openness implies that around a given right-hand side \(y\), the set of possible values \(T(U)\) for neighborhoods \(U\) is open in \(Y\). This yields robustness: small perturbations in \(y\) correspond to solvable perturbations in \(x\) lying within controlled neighborhoods.

4.4 Using openness to prove stability of solution sets

Another consequence is the stability of preimages of open sets:

If \(T\) is open and \(V\subseteq Y\) is open, then \(T^{-1}(V)\) can be expressed in terms of open sets in \(X\) by exploring neighborhoods whose images cover \(V\).
In variational problems, this reasoning underlies continuity of solution operators under suitable conditions, ensuring that small changes in data lead to solution sets that do not collapse abruptly.

These ideas are frequently expressed through derived estimates rather than direct topological arguments, but openness provides the guiding intuition.

5 Technical tools and proof patterns

5.1 Using neighborhood and basis arguments

Many proofs reduce openness to statements about a neighborhood basis.

A standard approach:

Choose a basis (or a convenient neighborhood family) \(\mathcal{B}\) at each point of \(X\).
Show that the image of each basis element is open in \(Y\).
Conclude openness for arbitrary open sets via union properties.

This pattern is especially effective in metric and topological vector spaces where balls or standard neighborhood sets form natural bases.

5.2 Handling images of unions and intersections of open sets

Open sets are unions of basis elements, so openness of \(f\) is compatible with unions in a straightforward way. However, images do not generally commute with intersections: typically, \[ f(U_1\cap U_2)\neq f(U_1)\cap f(U_2) \] so proofs that rely on intersections must be handled carefully.

A common technique is to avoid direct intersection statements and instead use local neighborhood arguments: represent a neighborhood by containment inside one of several open sets where the relevant property can be verified.

5.3 Metric-space formulations (open balls and continuity of neighborhoods)

In metric spaces, one can express openness using balls. For \(x\in X\), openness of \(f\) implies that for each \(\varepsilon>0\) with \(B_X(x,\varepsilon)\) open, the set \[ f(B_X(x,\varepsilon)) \] is open in \(Y\). While this is essentially the definition, metric settings allow further geometric estimates. In linear analysis, norm inequalities can show that images of balls contain balls around the image point, which directly proves openness.

5.4 Baire category and completeness requirements

In the Banach-space context, proofs of open mapping theorems often use the Baire category theorem. Completeness plays a central role: Baire’s theorem can fail in incomplete spaces, and accordingly the conclusion that surjective bounded linear maps are open may fail without completeness.

The typical strategy is to show that certain unions of images of scaled neighborhoods must cover a ball in \(Y\), forcing the map to have a quantitative surjectivity behavior that implies openness.

6 Variants and generalizations

6.1 Locally open and partially open maps

A map may be locally open without being globally open: each point in \(X\) has a neighborhood whose image is open, but images of all open sets might not be open. Local openness is common in geometric settings where maps look like covering maps or immersions around each point.

Another generalization is partial openness, where openness holds for a restricted class of open sets (for example, those belonging to a chosen basis or those with special geometric form).

6.2 Relatively open maps (subspace context)

When maps land in a subspace \(Z\subseteq Y\), one may ask whether \(f(U)\) is open in \(Z\) rather than in the whole \(Y\). This distinction matters because a set can be open in \(Z\) while not being open in \(Y\). Many examples involving embeddings and restrictions become clear once this relative viewpoint is adopted.

6.3 Graph-theoretic and categorical perspectives

Categorically, openness can be interpreted as a property of how morphisms interact with the lattice of open sets. In particular, the condition “images of opens are open” is a statement about mapping between frames (or locales) in a way that resembles preservation of structure.

Graph-theoretic analogies sometimes describe openness as a form of “no-gap propagation” of adjacency: open neighborhoods around points are carried to open neighborhoods in the target. These perspectives are not always formal in elementary topology, but they help explain why openness is considered a strong regularity condition in various abstractions.

6.4 Approximate openness and qualitative versions

In some analytic contexts, one studies weaker forms than strict openness, such as:

images that contain neighborhoods up to constants,
sets whose interior is nonempty rather than open,
or conditions that guarantee “thick” images without full openness.

These qualitative variants are useful when exact topological openness is too strong to prove, but geometric non-degeneracy is still needed for stability or existence arguments.

7 Worked examples and applications

7.1 Projection maps and their openness conditions

For a product \(X\times Y\) with product topology, the projection \(\pi_Y(x,y)=y\) is open because images of basic open sets \(U\times V\) are just \(V\). More generally, projections from products of topological spaces behave well because basic opens factor.

If the topology on the product is altered (for example by using a different product-like topology), openness of projections may no longer hold, underscoring that openness depends on the specific topological structures.

7.2 Dilations, translations, and affine maps in analysis

In Euclidean spaces, translations and dilations are homeomorphisms and hence open. An affine map \(x\mapsto Ax+b\) is open precisely when the linear part \(A\) is open onto its image; in particular, if \(A\) is surjective, the affine map is open onto the whole codomain.

In normed spaces, affine maps that are continuous and surjective are typically open under appropriate completeness assumptions, with proofs often reducing to the linear case.

7.3 Nonlinear examples: when maps are (not) open

Nonlinear maps can be open in some regions and fail globally.

A map can be continuous and locally open but still not map every global open set to an open set if different pieces interact in the image.
If a nonlinear map collapses an open region onto a set with empty interior, openness fails even if the map is well-behaved on each small neighborhood.

These examples show that verifying openness usually requires global consideration of how open sets spread under the map.

7.4 Practical checklist for verifying openness

A common workflow is:

Confirm the topology sources. Ensure the domain and codomain topologies are the intended ones (subspace vs. ambient matters).
Reduce via a basis. Test openness on a basis of the domain.
Use geometric neighborhood containment. In metric or normed settings, try to show images of balls contain balls.
Check global surjectivity when using linear theory. For linear operators between Banach spaces, surjectivity is the key hypothesis for standard theorems.
Watch for dimension collapse. If the image lies in a lower-dimensional subset or has empty interior, openness is unlikely.

8 Common misconceptions and pitfalls

8.1 Confusing open maps with continuous maps

Openness and continuity are different. Continuity concerns preimages of open sets being open; openness concerns images of open sets being open. A map can satisfy one without the other.

8.2 Open vs. surjective vs. injective

Openness does not automatically imply injectivity or surjectivity:

An open map may fail to be injective (many-to-one behavior can still preserve openness).
An open map may fail to be surjective if parts of \(Y\) never appear as images of open sets.

Conversely, surjectivity alone does not guarantee openness outside the special settings where open mapping theorems apply.

8.3 Dependence on the chosen topology

Openness is sensitive to the topology on \(X\) and \(Y\). Refining the topology on \(X\) generally makes it harder to keep images of open sets open, while coarsening can make openness easier. Similarly, changing the codomain topology changes which subsets are considered open, affecting whether \(f(U)\) is open.

8.4 Misreading “open mapping” versus “open set” terminology

Students sometimes misread “open mapping” as if it referred to mappings between open subsets only. The definition is about how the map treats *all* open sets in the domain, not about restricting the map’s domain or codomain to open parts.