Few weeks ago, I read a book that define Ring Localization by using Universal Property (or Universal Mapping Problem). This is not the first time I see definition of Localization of a Ring. However when it is defined by using Universal Property, it seems a lot less intuitive than what I experienced.
This is also not the first time I see some mathematical objects are defined by using Universal Property. Among the first definitions I see, there are Projective and Injective modules, Tensor product, Flat modules, product, pullback, pushout etc. I have encountered some of these definitions before and nearly all of them (if not all) defined in ways that, in my opinion, more intuitive.
This case both on Localization example and other examples makes me wonder why we even bother to define those objects via Universal Property. Speaking from Category Theory point of view, this approach (using Universal diagram etc) is of course the one most suitable and gives more generality. However, I think only being the most suitable is not enough reasons to do this — as my own view is that it is okay to use harder approach if this aid understanding.
I therefore tried to make sense of this approach by looking for another justifications. The first visit to wikipedia resulted in:
|“||Before giving a formal definition of universal properties, we offer some motivation for studying such constructions.
Also, I would like to add another remark to the second bullet. If we know one easy representation of the object, by using the isomorphic argument, we can say something like this: “Let x be an element of … (an object defined by Universal Property) then, x can be written as … (the easy representation)”.
By then, I was almost satisfied with wikipedia’s explanation.But then I discussed with two friends, and one of them argues that “by defining an object by using its Universal Property, we practically shift our attention from the object itself to object in diagram (another object or morphism)”. He gave example that he found in a paper by saying “When the tensor product of modules A and B are 0? Apparently, it only happens if all bilinear morphism from Cartesian product of A and B are 0 map”. Now that’s a revelation, by using Universal Property definition of tensor product we shift our focus out of A and B to their bilinear morphism .. very neat 😀
Category Theory seems to get brighter 🙂