Talk:Variety (universal algebra)

Someone please define subvariety

In this entry, the word "subvariety" appears without explanation. Would someone please define it? Better yet, would someone create a brief entry for the term, and blue link it to this entry? It is fascinating to read that groups are not necessarily subvarieties of semigroups, but that abelian groups are all subvarieties of groups. Are Boolean algebras subvarieties of lattices? Of commutative monoids?202.36.179.65 09:20, 25 September 2007 (UTC)

I rewrote the definition, did that make it any clearer? Lattices lack the complement of operation of Boolean algebras, and commutative monoids have only one binary operation whereas lattices have two, so all three of these classes have different signatures and can't be sublattices of one another by that definition. --Vaughan Pratt (talk) 08:04, 25 March 2011 (UTC)

arity?

What is arity? LilHelpa (talk) 19:44, 3 March 2009 (UTC)

Arity is the number of arguments an operation takes. Constants are 0-ary, negation and inversion are 1-ary (unary), addition, subtraction, multiplication, and division are 2-ary (binary), and the triple product is 3-ary. JackSchmidt (talk) 20:12, 3 March 2009 (UTC)

Covariety is unexplained

The lead mentions "covariety" with a link to "coalgebraic structures", but which redirects to Coalgebra. However as far as I can tell a coalgebra is something that is not in the realm of universal algebra, since it supposes vector spaces over a field. So please explain, or remove the phrase. Marc van Leeuwen (talk) 16:29, 2 February 2011 (UTC)

I think what happened here is just the usual confusion between universal algebras and algebras over a field. Apparently there are coalgebras in both senses, related to each other just like the two kinds of algebras. The link was probably supposed to go to universal coalgebras, but the article is about coalgebras over a field. I agree it should be explained, but the best place for that is probably a new article universal coalgebra. There is a redirect to this article from covariety, so that explains why it's in the lead and bold. Hans Adler 17:15, 2 February 2011 (UTC)

Turns out the requisite article already exists, under the awkward name F-coalgebra, so I redirected the link accordingly.

Since associative coalgebras are a pretty special case of coalgebras (I've written a fair about coalgebras but never had occasion to write about associative coalgebras), perhaps the following moves should be made:

• Coalgebra --> Associative coalgebra
• F-coalgebra --> Coalgebra (with a hatnote to associative coalgebra)

The associative coalgebra crowd could reasonably argue that the name belongs to them since they had it first. However it may be worth proposing at those pages to see what people think. --Vaughan Pratt (talk) 15:52, 25 March 2011 (UTC)

Agreeing on ambiguous terminology will always be difficult. I'm pretty sure the coalgebra people will object to moving to Associative coalgebra, and one good reason for that is that associativity is not the heart of the matter (although the current article assumes that). Their term "coalgebra" is based on dualizing either algebra (ring theory) or algebra over a field, so the essential point is that a coalebra is an additional structure defined on a module or vector space. But as far as I could see the examples at F-coalgebra never mention a ring or a field. I agree though that the name "F-coalgebra" is awkward (what to do if the functor is named G?). Although I'm no expert in either field, it seems that F-coalgraba is something dual to universal algebra, so why not something like "Universal coalgebra" or "coalgebraic variety"? As for the current coalgebra, one could propose a move to coalgebra over a field to distribute the pain fairly, but I'm not sure people will like that either. Of course the real error was to let the term algebra become as ambiguous as it is in the first place. Marc van Leeuwen (talk) 12:05, 26 March 2011 (UTC)

One could argue that "coalgebra" was more appropriately applied to locales than to "co-operations" of the form X → F(X) as dual to operations of the form F(X) → X, since these are special cases of the more general form F(X) → G(X), whereas there is no comparable common generalization of frames and locales. An equational theory presented as the commutative diagrams in a category C can accommodate any mixture of all of these notions simultaneously, regardless of whether C is an ordinary category with discrete homsets or an additive category whose homsets are abelian groups (in which case the endomorphisms of any object form a ring). From this perspective coalgebras and F-coalgebras are different facets of the same very general subject having in common the operation form X → F(X), with F(X) = X ⊗ X in some additive category C in the case of coalgebras. --Vaughan Pratt (talk) 00:44, 27 March 2011 (UTC)