free vector space over a set
In this entry we constructthe free vector space over a set, orthe vector space generated by a set [1].For a set , we shall denote this vector space by .One application of this construction is given in [2],where the free vector space is used to define the tensor product
formodules.
To define the vector space , let us first define as aset. For a set and a field , we define
In other words, consists of functions that are non-zero onlyat finitely many points in .Here, we denote the identity element in by , andthe zero element
by .The vector space structure for is defined as follows. If and arefunctions in , then is the mapping .Similarly, if and , then is the mapping . It is not difficult tosee that these operations are well defined, i.e., both and are again functions in .
0.0.1 Basis for
If ,let us define the function by
These functions form a linearly independent basis for , i.e.,
(1) |
Here, the space consists of allfinite linear combinations of elements in .It is clear that any element in is a member in .Let us check the other direction. Suppose is a member in .Then, let be the distinct points in where is non-zero.We then have
and we have established equality in equation 1.
To see that the set is linearly independent,we need to show that its any finite subset is linearly independent.Let be sucha finite subset, andsuppose for some. Since the points are pairwise distinct, itfollows that for all . This shows that the set is linearly independent.
Let us define the mapping , .This mapping gives a bijection between and the basisvectors . We can thus identify thesespaces. Then becomes a linearly independent basis for .
0.0.2 Universal property of
The mapping is universal inthe following sense. If is an arbitrary mapping from to avector space , then there exists a unique mapping such that the below diagram commutes:
Proof. We define as the linear mapping thatmaps the basis elements of as .Then, by definition, is linear. For uniqueness,suppose that there are linear mappingssuch that .For all , we then have .Thus since both mappings are linear andthe coincide on the basis elements.
References
- 1 W. Greub,Linear Algebra
,Springer-Verlag, Fourth edition, 1975.
- 2 I. Madsen, J. Tornehave,From Calculus to Cohomology
,Cambridge University press,1997.