Dear Tom, I really appreciate your help, but If I could ask the perfect question I probably would already know the answer... My example should not prove anything, instead they collectively show, that I am missing something. And it is not the fact, that "pure f does not depend on f." If, however, this makes all the difference, I have to ask, why was plausability and looking at the actual definition (not just the types) not important for the other examples. But I think my problem lies somewhere else. Maybe all would become evident, if I knew the rigorous definition of "A is more general than B" in this context. Especially when A is a class of type, that takes two arguments (i.e. Unit and Arrow) and B for ones, that takes only one (like Monad, Pure,..) Thanks again! Johannes

On Tue, May 28, 2013 at 11:11 PM, Tom Ellis <tom-lists-haskell-cafe-2...@jaguarpaw.co.uk> wrote: > On Tue, May 28, 2013 at 09:09:48PM +0200, Johannes Gerer wrote: >> What about these two very simple type classes. Are they equivalent? > [...] >> class Pointed f where >> pure :: a -> f a >> >> class Unit f where >> unit :: f a a >> >> newtype UnitPointed f a = UnitPointed f a a >> instance Unit f => Pointed (UnitPointed f) where >> pure f = UnitPointed unit >> >> newtype Kleisli f a b = Kleisli (a -> f b) >> instance Pointed f => Unit (Kleisli f) where >> unit = Kleisli pure > > This is implausible, since "pure f" does not depend on "f". > > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe@haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe _______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe