← Latest papers
🔢 mathematics

A taxonomy of categories for relations

This paper provides a modern, organized taxonomy of categories that abstract the structural properties of relations, including their enriched versions and their characterization as Kleisli categories of symmetric monoidal monads.

Original authors: Cipriano Junior Cioffo, Fabio Gadducci, Davide Trotta

Published 2026-05-13
📖 5 min read🧠 Deep dive

Original authors: Cipriano Junior Cioffo, Fabio Gadducci, Davide Trotta

Original paper dedicated to the public domain under CC0 1.0 (http://creativecommons.org/publicdomain/zero/1.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to organize a massive library of different types of "connections" between things. In mathematics and computer science, we often study functions (where one input leads to exactly one output) and relations (where one input can lead to many outputs, or none at all).

Over the last few decades, mathematicians have invented dozens of different "rulebooks" (called categories) to describe how these connections work. The problem is that these rulebooks often have different names, even though they are describing very similar ideas. It's like having a library where "Apples" are called "Red Fruits," "Oranges" are called "Citrus," and "Bananas" are called "Yellow Curves," but no one has a map showing how they all fit together.

This paper, "A Taxonomy of Categories for Relations," by Cioffo, Gadducci, and Trotta, is essentially a master map or a periodic table for these connection rulebooks. Here is a simple breakdown of what they did:

1. The Building Blocks: Copying and Discarding

To understand their map, you first need to understand two basic actions that happen when things interact:

  • Copying (The "Share" Action): Imagine you have a document. You can make a copy of it. In math terms, this is taking one thing and turning it into two identical things.
  • Discarding (The "Garbage" Action): Imagine you have a document and you throw it in the trash. You don't need to know what was on it anymore; it just disappears.

The authors realized that almost every "connection rulebook" in the literature is built by deciding which of these two actions are allowed, and whether they follow strict rules (like "you must always be able to copy") or loose rules (like "you can copy, but maybe not always").

2. The "GS-Monoidal" Core

The authors introduce a central concept they call GS-monoidal categories. Think of this as the "Swiss Army Knife" of connection rulebooks.

  • GS stands for Garbage and Share.
  • If a rulebook allows you to copy things, it has "Share" structure.
  • If it allows you to throw things away, it has "Garbage" structure.
  • If it allows you to do both, it's a GS-monoidal category.

They show that many famous concepts in math and computer science are just specific versions of this Swiss Army Knife:

  • Markov Categories: These are rulebooks for probability. They are like "Garbage" rulebooks where you must be able to throw things away (representing the idea that probabilities must sum up to 1).
  • Restriction Categories: These are rulebooks for partial functions (where a function might fail or not exist). They are like "Share" rulebooks where you can copy things, but only under certain conditions.
  • Cartesian Categories: These are the standard "total functions" we learn in school. They are the most rigid version, where you can always copy and always discard perfectly.

3. The "Kleisli" Machine

The paper also looks at a specific mathematical machine called a Kleisli category.

  • The Metaphor: Imagine you have a standard factory (a category) that makes widgets. Now, imagine you add a "wrapper" or a "special effect" to the factory (called a Monad). The Kleisli category is the new factory that produces "wrapped widgets."
  • The Discovery: The authors prove that if you take a "Garbage/Share" factory and wrap it with a specific type of special effect, the new factory still keeps the Garbage/Share rules.
  • Why it matters: This helps mathematicians know that if they build a complex system using these wrappers, they don't lose the fundamental properties of copying and discarding. It's like saying, "If you put a protective case on a Swiss Army Knife, it's still a Swiss Army Knife."

4. The "Enriched" Version (Adding a Ladder)

Finally, the paper looks at a more complex version where the connections aren't just "yes/no" but have a ranking or order (like a ladder).

  • The Metaphor: In a normal rulebook, two connections are either the same or different. In this "enriched" version, one connection can be "less than" or "better than" another.
  • They show that even with this extra ladder of ranking, the same "Garbage/Share" rules still apply, just with a few extra inequalities (like saying "Copying is at least as good as doing nothing").

The Big Picture

The authors didn't invent new "magic" connections. Instead, they took a chaotic library of existing ideas and organized them into a clean, logical family tree.

  • They showed that many different names (Markov, Restriction, Affine, etc.) are actually just different combinations of "Copying" and "Throwing Away."
  • They showed how these structures behave when you apply mathematical "wrappers" (monads) to them.
  • They provided a single, unified language (using "string diagrams," which look like circuit boards) to talk about all of them at once.

In short, this paper is a translator and organizer that helps researchers stop getting confused by different names and start seeing the underlying unity in how mathematical relations work.

Drowning in papers in your field?

Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.

Try Digest →