Fibers and Sections

Patrick Reany

1 October 2025

Note: This chat I had with Copilot is likely edited for relevance and brevity. It is also likely that I will not take the time to highlight all the words that Copilot highlights. Further, in the interest of being didactic, I might arbitrarily highlight an occasional word or phrase that Copilot did not. (Naturally, I get to correct my spelling, grammar, and other minor errors for free.)

Patrick:
Let $A$ and $B$ be sets and let $f:A \to B$, where $f$ is a function. For each $y \in B$ (which is in the image of $f$), the set $f^{-1}(y)$ is a subset of $A$. I have a source which calls this subset the 'fiber' of $y$. Is this correct usage or an abuse of notation?

Copilot:
Yes, this is correct usage, not an abuse of notation.

What Is a Fiber?

Given a function \( f: A \to B \), and an element \( y \in \text{im}(f) \subseteq B \), the fiber over \( y \) is defined as: \[ f^{-1}(y) = \{ a \in A \mid f(a) = y \} \] This is the preimage of the point \( y \), and it's a subset of \( A \). The term "fiber" is widely used in mathematics to describe this set --- especially in contexts like:

• Topology and geometry: In fiber bundles, each point in the base space has a fiber above it.
• Algebraic geometry: Fibers of morphisms help analyze the structure of varieties.
• Category theory: Fibers appear in slice categories and pullbacks.
• Set theory and analysis: Even in elementary settings, fibers help describe the structure of functions.

Why It's Not Abusive

Calling \( f^{-1}(y) \) the fiber of \( y \) is not just convenient --- it reflects a deep geometric and categorical intuition: the idea that a function "projects" elements of \( A \) onto \( B \), and the fiber over \( y \) collects all elements that land there.

So your source is using the term in a mathematically sound and conceptually rich way. If anything, it's inviting you into a broader landscape where functions are seen as bundles of fibers, each with its own internal structure.

Would you like to explore how this idea generalizes to fiber bundles or sheaves? Or how it relates to equivalence relations and quotient sets?

Patrick:
So, is the name 'fiber' just a convenient way to describe any association of a collection of things to a point in some 'base' space?

Copilot:
Yes --- and beautifully put, Patrick. The term fiber is more than just convenient; it's a conceptual metaphor that elegantly captures the idea of associating a collection of things to each point in a base space. This association can be purely set-theoretic, geometric, algebraic, or even categorical, depending on context.

What Makes a Fiber a Fiber?

At its core, a fiber is:

• A collection of elements (often forming a set, space, or structure)
• Indexed by or attached to a point in some base space
• Often arising from a function, projection, or morphism that "collapses" the fiber into the base point

This abstraction shows up across mathematics:

Philosophical Flavor

Calling these preimages "fibers" invites a geometric or architectural intuition: imagine the base space as a loom, and the fibers as threads running vertically, each one carrying its own internal structure. The function or morphism then acts like a projection, flattening the tapestry into the base.

So yes --- "fiber" is a flexible, evocative term that helps unify diverse mathematical ideas under a shared metaphor. Want to riff on how this metaphor evolves in higher category theory or homotopy type theory?

Patrick:
Then in addition to the description I already gave for a 'fiber', it also needs a formal means to project its contents onto its base point, right?

Copilot:
Exactly --- and that's a crucial insight. A fiber isn't just a set floating in abstraction; it's meaningful only in relation to a projection or structure-preserving map that anchors each fiber to a point in the base space.

The Role of the Projection To formally define a fiber, you need:

1. A total space \( E \) --- the domain where all the fibers live.
2. A base space \( B \) --- the set of points over which fibers are defined.
3. A projection map \( \pi: E \to B \) --- which assigns each element of \( E \) to a point in \( B \).

Then, for each \( b \in B \), the fiber over \( b \) is: \[ \pi^{-1}(b) = \{ e \in E \mid \pi(e) = b \} \] This projection is what gives the fiber its identity. Without it, the fiber is just a set --- with it, it becomes a structured slice of the total space, tied to a specific location in the base.

Examples Across Domains

• Set theory: \( f: A \to B \), fiber over \( y \in B \) is \( f^{-1}(y) \).
• Fiber bundles: \( \pi: E \to B \), where each fiber \( \pi^{-1}(b) \) is a vector space or manifold.
• Sheaves: The stalk over a point is the fiber of the sheaf at that point.
• Categories: In a fibration, the fiber over an object is a subcategory mapped to it.

Philosophical Echo

Your phrasing --- "a formal means to project its contents onto its base point" --- captures the essence of what makes a fiber relational. It's not just a collection; it's a contextualized collection, defined by its place in a larger structure.

Would you like to explore how this projection idea evolves in dependent type theory or higher category theory, where fibers become entire homotopy types or $\infty$-groupoids?

Patrick:
Then a "section" would be a choice of $\xi$ in $f^{-1}(y)$ for every point $y \in B$, such that $f(\xi) = y$, right?

Copilot:
Yes --- that's precisely the idea of a section, and you've captured it elegantly.

Formal Definition of a Section

Given a function (or more generally, a projection) \( f: A \to B \), a section is a map: \[ s: B \to A \] such that: \[ f(s(y)) = y \quad \text{for all } y \in B \] In other words, \( s(y) \in f^{-1}(y) \), so the section picks out one element from each fiber over \( y \in B \). It's a kind of right inverse to \( f \), though not necessarily globally invertible.

Intuition

• Think of \( f \) as collapsing structure: it sends many elements of \( A \) to a single point in \( B \).
• A section \( s \) reverses that collapse --- it selects a representative from each fiber.
• If you imagine \( f \) as a projection from a bundle of threads onto a base, then a section is like choosing one thread from each bundle.

In Broader Contexts

• In topology, a section of a fiber bundle is a continuous map that selects a point in each fiber.
• In sheaf theory, a section over an open set is a function assigning data to each point in that set, compatible with the sheaf structure.
• In category theory, sections are morphisms that split epimorphisms.

So yes --- your description is spot-on. A section is a systematic choice of \( \xi \in f^{-1}(y) \) for each \( y \in B \), such that \( f(\xi) = y \). Want to explore when such sections exist, or how they relate to the Axiom of Choice?

Patrick:
Are the points in $B$ that are not in the image of $f$ included in these fibers over the base points, or are they left out, being trivially empty?

Copilot:
They're left out --- or more precisely, their fibers are trivially empty.

Fiber Over a Point Not in the Image

Given a function \( f: A \to B \), for any \( y \in B \), the fiber over \( y \) is: \[ f^{-1}(y) = \{ a \in A \mid f(a) = y \} \] Now, if \( y \notin \text{im}(f) \), then there are no elements \( a \in A \) such that \( f(a) = y \). So: \[ f^{-1}(y) = \emptyset \] This is still a valid fiber --- it's just the empty set. So technically, every point in \( B \) has a fiber, but only those in the image of \( f \) have non-empty fibers.

Conceptual Note

• In set theory, this is straightforward: the preimage of a point not in the image is empty.
• In topology or geometry, the empty fiber might signal a discontinuity or a singularity.
• In category theory, the fiber over an object might be empty if no morphisms land there.

So yes --- points in \( B \) not in the image of \( f \) are associated with empty fibers. They're part of the formal structure, but they carry no content. Want to explore how this plays out in sheaf theory or logic, where empty fibers can have deep interpretive meaning?

Patrick:
If we restrict $f$ to just the section points of $A$ then $f$ is 1-1. But is this a particularly useful fact, or just barely interesting for most applications?

Copilot:
It's more than barely interesting --- but its usefulness depends on context. Let's unpack it.

What You've Observed

If \( f: A \to B \) is a function, and \( s: B \to A \) is a section (i.e., \( f(s(y)) = y \) for all \( y \in B \)), then the restriction of \( f \) to the image of \( s \), denoted \( \text{im}(s) \subseteq A \), is injective: \[ f|_{\text{im}(s)}: \text{im}(s) \to B \] Why? Because if \( f(s(y_1)) = f(s(y_2)) \), then \( y_1 = y_2 \), so \( s(y_1) = s(y_2) \), proving injectivity on that subset.

Why This Matters

This fact is structurally significant in several areas:

• Set theory & logic: It's a constructive way to build a right inverse, which is closely tied to the Axiom of Choice. If every surjection has a section, then choice holds.
• Category theory: Sections split epimorphisms. The injectivity of the restricted map reflects the idea that the section embeds the base into the total space.
• Topology & geometry: In fiber bundles, a section gives a "slice" through the bundle. Injectivity ensures that this slice doesn't fold or overlap.
• Algebra: In group theory or module theory, sections often correspond to splittings of exact sequences, and injectivity is crucial for decompositions.

Philosophical Flavor

You're noticing that a section doesn't just pick representatives --- it reconstructs the base faithfully. That’s why the restricted map is injective: it's a mirror of the base inside the total space.

So yes, it's useful --- especially when you care about reversibility, structure preservation, or constructive embeddings. Want to explore how this plays out in sheaf cohomology or dependent type theory?

Afterword: