Distributive category

From The Right Wiki
Jump to navigationJump to search

In mathematics, a category is distributive if it has finite products and finite coproducts and such that for every choice of objects A,B,C, the canonical map

[idA×ι1,idA×ι2]:A×B+A×CA×(B+C)

is an isomorphism, and for all objects A, the canonical map 0A×0 is an isomorphism (where 0 denotes the initial object). Equivalently, if for every object A the endofunctor A× defined by BA×B preserves coproducts up to isomorphisms f.[1] It follows that f and aforementioned canonical maps are equal for each choice of objects. In particular, if the functor A× has a right adjoint (i.e., if the category is cartesian closed), it necessarily preserves all colimits, and thus any cartesian closed category with finite coproducts (i.e., any bicartesian closed category) is distributive.

Example

The category of sets is distributive. Let A, B, and C be sets. Then

A×(B⨿C)={(a,d)aA and dB⨿C}{(a,d)aA and dB}⨿{(a,d)aA and dC}=(A×B)⨿(A×C)

where ⨿ denotes the coproduct in Set, namely the disjoint union, and denotes a bijection. In the case where A, B, and C are finite sets, this result reflects the distributive property: the above sets each have cardinality |A|(|B|+|C|)=|A||B|+|A||C|. The categories Grp and Ab are not distributive, even though they have both products and coproducts. An even simpler category that has both products and coproducts but is not distributive is the category of pointed sets.[2]

References

  1. Taylor, Paul (1999). Practical Foundations of Mathematics. Cambridge University Press. p. 275.
  2. F. W. Lawvere; Stephen Hoel Schanuel (2009). Conceptual Mathematics: A First Introduction to Categories (2nd ed.). Cambridge University Press. pp. 296–298. ISBN 978-0-521-89485-2.

Further reading