Not signed in (Sign In)

Not signed in

Want to take part in these discussions? Sign in if you have an account, or apply for one below

  • Sign in using OpenID

Discussion Tag Cloud

Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.

Welcome to nForum
If you want to take part in these discussions either sign in now (if you have an account), apply for one now (if you don't).
    • CommentRowNumber1.
    • CommentAuthorBen_Sprott
    • CommentTimeJul 18th 2020
    • (edited Jul 18th 2020)

    Naturalness in Category Theory is a difficult notion to pin down. Here is some discussion about naturalness.

    I try to study and generate applications of Category Theory to the natural world. I believe that naturalness has been a guiding principle in my work over the years. Another principle that has guided me is Occams’s razor. I get the feeling they may be related, but I have never seen this written about. Does anyone have any thoughts on this?

    I have been developing ideas about science couched in the language of monads. They seemed rather natural to me. I am now discovering sketches, but have a very limited understanding of both of these. They both seem like a way to have natural presentations of theories or analysis of ideas. Is one more natural than the other? I feel that monads are more natural, but that might just be because I don’t understand sketches very well. I think monads are more natural because there is this effect that, when you are looking for a given monad, it may only exist on a particular category. It forces you to accept the particular site where everything that ought to be true, is true. Cleverness, or brittle constructions are not as useful.

    I suppose the counter argument could be made that sketches work by choosing the kinds of limits you are going to use and then you can achieve various presentations of theories given that kind of limits you chose. So, you work in the universe or logic of particular limits and all things that ought to be true with those limits are true.

    It has been put forth that monads and sketches are not related, in that they don’t present the same kinds of things (or theories). Taking a look at Barr and Well’s CT for computing science text, the first example they give is the “free monoid construction”. How is that any different from the list monad, who’s category of algebras is the category of monoids? Also, we know that you can generate monads from sketches. They are clearly related, and I would argue, presentations of the same things.

    Here we see a publication where the authors assert the following

    “Lawvere theories and monads have been the two main category theoretic formulations of universal algebra”

    I believe Lawvere Theories are just sketches restricted to finite products.

    The fact that sketches tend to be more like complicated constructions, would suggest to me that they are less natural than monads.

    Why is this all important? Look at this quote in Spivak’s latest paper

    “Previous work has shown how to allow collections of machines to reconfig- ure their wiring diagram dynamically, based on their collective state. This notion was called “mode dependence”, and while the framework was compositional (forming an operad of re-wiring diagrams and algebra of mode-dependent dynamical systems on it), the formulation itself was more “creative” than it was natural.”

    Naturalness mattered to the tool he choses in his paper, namely the category of polynomial endofunctors on set and their monads

    If you read this book by Hosseenfelder, you see that what physics may be in need of now is a new guiding principle, other than the standard understanding of beauty. I am proposing that the new principle which we should be looking for is the categorical notion of naturalness.

  1. Hi Ben, interesting question! I wouldn’t criticise sketches for their complexity; it is true that writing them down in full detail is rather tedious, but this is more of a practical problem than a philosophical one! I rather like sketches myself; one particularly nice aspect is the fact that they isolate exactly what category theoretic structure (doctrine, if you like) is required in order to have a model. Perhaps for this reason, as you probably know, they have interesting theoretical applications, for instance with regard to accessiblity/local presentability. Another important point, which you touched upon, is that sketches are much more general: categories of algebras for monads correspond to models of a very restricted kind of sketch.

    Both sketches and monads can be thought of as ’purely category theoretic’ notions of theory, so in that sense are more or less equally ’natural’, although I typically actually tend to think of sketches as ontologically more ’graph-like’, if you will, than ’category-like’. The great power of a monad, as I see it, is that it is a ’higher level’ notion, i.e. at a greater level of abstraction, than a sketch, and also much more concise. This means that one can typically apply more powerful category theoretic machinery to work with it.

    I don’t know if any of this is helpful!

    • CommentRowNumber3.
    • CommentAuthorBen_Sprott
    • CommentTimeJul 18th 2020
    • (edited Jul 18th 2020)

    Thank you Richard!

    I get the feeling that researchers may start to try to cast physics in a foundation of either monads or sketches. I have spent a few months trying to do this with sketches, and many years for monads. The fact that sketches, as you have pointed out, are more graph like, than category like, it may have greater appeal to those attempting a discrete causal approach to physics. I am thinking this might be a mistake and, I guess, I am addressing it early.