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.
    • CommentAuthortomr
    • CommentTimeFeb 19th 2013
    • (edited Feb 19th 2013)
    I am making some steps in coalgebra studies (they are tools for investigation of state transition systems (STS) and STS have immense applications in computer science, robotics and so on), but I am taken aback by quote from J. J. M. M. Rutten (one who has written possibly the best reviews and tutorials about coalgebra applications in computer science):
    "Deep insights about groups are not obtained by studying universal algebra. Nor
    will universal coalgebra lead to di cult theorems about (speci c types of) systems.
    Like universal algebra, its possible merit consists of the fact that it tidies up
    a mass of rather trivial detail, allowing us to concentrate our powers on the hard
    core of the problem"

    After reading some articles I should start to believe this quote, but the question is - is there possible to consider some more concrete (more refined) structures than coalgerbas and arrive at less general (more practical) results. Judging from the algebra courses, on can expect (at least by playing with words) that cogroups (and their representations) could be really fruitful objects of research. But there are so little literature about them. Why is so? I can guess - maybe all the results for cogroups can be trivially achived from the group theory? Maybe mathematicians don't see the perspective in this field and so on?

    From reading ncatlab topics I can understand that the term "cogroup" is used in two meanings - as cogroup object and simply as cogroup. I am more interested in the later - cogroup as special case of coalgebra.
    • CommentRowNumber2.
    • CommentAuthortomr
    • CommentTimeFeb 19th 2013
    • (edited Feb 19th 2013)
    The only book about this topic that I am aware is "Cogroups and Co-Rings in Categories of Associative Rings", but maybe there is something else?
    • CommentRowNumber3.
    • CommentAuthorTodd_Trimble
    • CommentTimeFeb 19th 2013

    I don’t think the quote of Rutten is so much a hard truth as it is a general piece of wisdom. It reminds me very much of Freyd’s quip, “Perhaps the purpose of categorical algebra is to show that which is trivial is trivially trivial,” which is glossed here at the nLab. But I can think of cases where the soft methods of category theory and universal algebra lead directly to solutions of problems that some people have apparently found hard.

    Anyway, you might expand your search terms to include things like “Hopf algebras” or “comonoids”.

    In all such discussions, a key question is the background in which we define comonoids and cogroups: what is the monoidal product relative to which we define them? For example, if we use cartesian product as the monoidal product, then a famous exercise is that every object carries exactly one comonoid structure (which makes them uninteresting in some sense). Often, when people speak of cogroups, they dualize the notion or group (which uses cartesian product) so that the relevant monoidal product is the coproduct. However, in the category of sets (or other categories where the initial object is strict, e.g., a cartesian closed category or an extensive category such as SetSet or TopTop, the only cogroup is the initial object! For this reason, you basically never hear about cogroups except in the context of cogroup objects! For example, the spheres famously have cogroup object structures in the category of pointed topological spaces.

    I don’t have a great answer why cogroup objects etc. are less studied than their dual counterparts, other than to say the collective consciousness has a lot more experience with induction than co-induction, or that some concepts as applied to concrete structures seem to be harder for human brains than their dual counterparts, e.g., subobjects are usually easier in concrete settings than quotient objects.

    • CommentRowNumber4.
    • CommentAuthorTim_Porter
    • CommentTimeFeb 20th 2013

    Perhaps one idea that is relevant is to look at the coalgebras that arise as models of modal logics. In that area one has also algebra type structures since on has an extended Stone duality theory. The algebras are Boolean algebras with operators, and the coalgebras various types of relational structure. The different relational structures do get studied in different ways, using different tools. For instance, equivalence relations are usually not thought of from a coalgebraic viewpoint, and they tend to be used rather than studied! They are examples of groupoids and that can be useful, and the underlying graph is a good intuitive way of handling them. Perhaps one needs to look at the covarieties of the relational structures / coalgebras rather than all of them.

    Finally groups are useful because they correspond to symmetries of objects. Cogroups need not be so useful, unless they arise in some area as ???cosymmetries???.

    • CommentRowNumber5.
    • CommentAuthortomr
    • CommentTimeMar 11th 2013
    Recent reading of excellent paper (preprint) "Modal Logics are Coalgebraic" confirms suggestion by Tim_Porter and also contradicts the quote of J. J. M. M. Rutten. As it is said in mentioned preprint:
    "In this way, one obtains e.g.
    algorithms for reasoning with coalgebraic logics that
    uniformly cover all the previously given examples. A
    surprisingly large body of results can be obtained at this
    high level of generality, indicating that coalgebras relate
    to modal logic at precisely the right level of abstraction."