Table of contents |

2 Examples 3 Properties |

Let **C** be a preadditive category.
In particular, morphisms in **C** can be added.

Given objects *A*_{1},...,*A*_{n} in **C**, suppose that we have:

- another object
*A*_{1}⊕ ··· ⊕*A*_{n}in**C**(the*biproduct*); - morphsims
*p*_{k}:*A*_{1}⊕ ··· ⊕*A*_{n}→*A*_{k}in**C**(the*projection morphisms*); and - morphisms
*i*_{k}:*A*_{k}→*A*_{1}⊕ ··· ⊕*A*_{n}(the*injection morphisms*).

- (
*i*_{1}_{°}*p*_{1}) + ··· + (*i*_{n}_{°}*p*_{n}) equals the identity morphism of*A*_{1}⊕ ··· ⊕*A*_{n}; *p*_{k}_{°}*i*_{k}equals the identity element of*A*_{k}; and*p*_{k}_{°}*i*_{k'}is the zero morphism from*A*_{k'}to*A*_{k}whenever*k*and*k*' are distinct.

Note that if we take *n* = 0 in the above definition, then only the first condition applies, and we have for the *nullary biproduct* an object *O* such that the identity morphism on *O* is equal to the zero morphism from *O* to itself.

Biproducts always exist in the category of abelian groups. In that category, the biproduct of several objects is simply their direct sum. The nullary biproduct is the trivial group. Biproducts exist in several other categories with direct sums, such as the category of vector spaces over a given field. But biproducts do not exist in the category of all groupss; indeed, this category is not even preadditive.

If a nullary biproduct exists and all binary biproducts *A*_{1} ⊕ *A*_{2} exist, then all biproducts whatsoever must also exist; this can be proved by mathematical induction.

Biproducts in preadditive categories are always both productss and coproducts in the ordinary category-theoretic sense; this is the origin of the term "biproduct". In particular, a nullary biproduct is always a zero object. Conversely, any finitary product or coproduct in a preadditive category must be a biproduct.

An *additive category* is a preadditive category in which every biproduct exists.
In particular, biproducts always exist in abelian categories.