The nucleus: Mining concepts from adjunctions
Year: 2020
Type: preprint
Abstract: Categories are many things to many people. We like to think of category theory as the study of certain types of grammar: those that generate the languages of mathematical constructions. This view seems to explain why categories persist in many parts of mathematics despite the ongoing objections that they do not help solve hard problems. Grammars do not help us tell better stories, but they help us tell stories better, and understand them better. Moreover, a language has grammar whether the speakers are aware of it or not.
A grammatical schema that generates many mathematical constructions is that of an adjunction. Lawvere noticed that all standard logical operations, quantifiers, and even the set-theoretic comprehension, can be viewed as adjunctions. Lambek observed that natural language grammars also decompose into certain adjunctions, spawning an entire branch of linguistics. Adjunctions are everywhere; all of category theory can be reconstructed in terms of adjunctions. If category theory is a grammar of mathematical constructions, then adjunctions are the one rule to rule them all. [snip]
Any adjunction induces a monad and a comonad, which respectively display its algebraic and coalgebraic (analytic, state-dependent) aspects. The nucleus construction induces a monad over adjunctions themselves (or equivalently over the induced monads, or comonads), which distills both aspects. It is tight because it is idempotent. The two sides of a nuclear adjunction completely determine one another, as algebras and coalgebras, yet they remain distinct, like brackets.
Source: arXiv (Cornell University)
Authors: Duško Pavlović, Dominic J. D. Hughes
Language: English
Topic: Advanced Algebra and Logic
Subfield: Computational Theory and Mathematics
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3016483071
