1 Definition and basic idea

A free vector space on a set is the vector space generated by that set in the most unrestricted way possible, subject only to the axioms of vector spaces. Its elements are not the original set elements themselves, but formal sums formed from them with coefficients from a chosen field. This construction is useful because it embeds an arbitrary set into a linear setting while preserving distinctness of the original elements as basis vectors.

1.1 Formal linear combinations

An element of a free vector space is a finite formal linear combination of symbols taken from the underlying set. Such an expression has the form \[ a_1 s_1 + a_2 s_2 + \cdots + a_n s_n, \] where each \(a_i\) is a scalar and each \(s_i\) is an element of the set. The coefficients determine how the symbols are weighted, and two combinations are considered equal only when corresponding coefficients match after like terms are combined.

1.2 Underlying set and basis elements

The original set serves as the source of basis symbols, not as a vector space already equipped with operations. Each element of the set is represented by a distinct basis vector in the free vector space. This makes the construction canonical: no extra relations are imposed among the generators, so the only dependencies arise from the axioms of vector addition and scalar multiplication.

1.3 Finite support condition

Only finitely many coefficients in any formal sum are allowed to be nonzero. This restriction ensures that addition and scalar multiplication are well defined without requiring convergence or infinite summation. The set of indices with nonzero coefficients is called the support of the vector, and it is finite for every element of a free vector space.

2 Construction

The free vector space is obtained by taking one basis symbol for each element of a set and then forming all finite linear combinations of those symbols. The field used for coefficients is fixed in advance, and the resulting vector space depends on that choice. The construction works for any set, whether finite or infinite.

2.1 From a given set

Starting from a set \(S\), one introduces a symbol \([s]\) or similar for each \(s \in S\). The vector space consists of finite sums of these symbols with coefficients from the chosen field. The operations are defined termwise and then extended by linearity.

2.1.1 Symbols as basis vectors

Each element of the set is treated as a formal generator. These generators are independent placeholders rather than numbers or vectors already present in a surrounding space. In the free vector space, the symbol for a given set element becomes a genuine basis vector, and distinct elements of the set correspond to distinct basis elements.

2.1.2 Vector addition and scalar multiplication

Addition is defined by combining like symbols and adding their coefficients. Scalar multiplication multiplies every coefficient by the scalar. Since all vectors have finite support, these operations produce another finite formal sum and satisfy the usual vector space axioms.

2.2 Examples of free vector spaces

Concrete examples help clarify how the construction behaves in different cases. The size of the basis is exactly the size of the original set, although the space itself may have many more vectors than the set has elements. Even a small generating set can produce a large family of formal combinations.

2.2.1 Free vector space on a finite set

If the set has \(n\) elements, the free vector space is naturally isomorphic to the \(n\)-dimensional vector space over the chosen field. Each element of the set corresponds to one standard basis vector. In this case, the free vector space is familiar as the space of coordinate vectors with \(n\) components.

2.2.2 Free vector space on an infinite set

If the set is infinite, the free vector space still consists of finite linear combinations, so each vector involves only finitely many generators. The dimension is equal to the cardinality of the set. This produces large vector spaces with a basis indexed by an infinite collection of symbols.

3 Universal property

The key feature of a free vector space is that it characterizes linear extension from arbitrary functions on the generating set. This property explains why the construction is called free: it imposes no relations beyond linearity. Many arguments about maps out of such spaces reduce to checking behavior on basis elements.

3.1 Extension of functions

Any function from the generating set into a vector space extends uniquely to a linear map from the free vector space. The extension is obtained by applying the function to each basis symbol and then distributing over the finite sum. This makes the free vector space a canonical tool for translating set-based data into linear-algebraic form.

3.2 Uniqueness of linear extensions

The extension is unique because a linear map is determined entirely by its values on a basis. Once the images of the generating symbols are fixed, the images of all finite linear combinations are forced. This uniqueness is central to the universal property and is often used to define maps indirectly.

3.3 Functorial viewpoint

The assignment sending a set to its free vector space can be viewed as a functor from sets to vector spaces. A function between sets induces a linear map between the corresponding free vector spaces by sending each basis element to the image of its generator. In this way, the construction is compatible with composition and identity maps.

4 Basis and dimension

The basis of a free vector space is built into the construction, rather than chosen afterward. This makes the space a standard example of a vector space with a specified basis. Its dimension is determined entirely by the size of the underlying set.

4.1 Canonical basis

The generators corresponding to elements of the original set form a distinguished basis. Because they arise directly from the set, this basis is often called canonical. The existence of such a preferred basis distinguishes free vector spaces from arbitrary vector spaces, which may have many possible bases but no preferred one.

4.2 Linear independence

The canonical basis is linearly independent. No nontrivial finite linear combination of distinct generators can equal zero unless all coefficients vanish. This independence reflects the absence of relations among the generators, which is precisely what makes the space free.

4.3 Dimension of the free vector space

The dimension of a free vector space equals the cardinality of the set on which it is based. For a finite set, this gives the usual finite dimension. For an infinite set, the dimension is the corresponding infinite cardinal, even though each individual vector still uses only finitely many basis elements.

5 Relationship to other algebraic structures

Free vector spaces are prototypes for several other free constructions in algebra. They show how a set can be transformed into an algebraic object that has a universal mapping property. Related notions include free modules, free abelian groups, and multilinear constructions.

5.1 Free modules

A free vector space is a special case of a free module over a field. In module theory, the coefficients come from a ring rather than a field, but the same idea applies: one takes formal finite linear combinations of basis symbols. The vector-space case is simpler because every nonzero scalar is invertible.

5.2 Free abelian groups

Free abelian groups are analogous constructions in additive group theory, where coefficients are integers rather than field elements. Like free vector spaces, they are generated by basis elements with no relations other than the group axioms. The comparison helps explain how linear algebra fits into a broader pattern of universal algebraic constructions.

5.3 Tensor products and multilinear maps

Free vector spaces are often used in the study of tensor products because they provide a convenient starting point for imposing multilinear relations. A multilinear map can be analyzed by first extending it linearly from generators and then factoring through suitable quotients. This approach turns complicated multilinear behavior into a problem about linear maps between free objects.

6 Applications

Free vector spaces appear throughout algebra whenever one needs to linearize a set or encode combinatorial data in vector-space form. They serve both as technical tools and as conceptual bridges between discrete and linear structures. Their universal property makes them especially useful in constructions and proofs.

6.1 Abstract linear algebra

In abstract linear algebra, free vector spaces provide the cleanest example of a vector space with a prescribed basis. They are used to build models, prove existence statements, and reduce general claims to computations on basis elements. Many proofs involving linear maps begin with a free construction and then apply a universal property.

6.2 Category-theoretic constructions

From a categorical perspective, the free vector space functor is left adjoint to the forgetful functor from vector spaces to sets. This adjunction encapsulates the idea that free vector spaces are the most economical way to impose linear structure on a set. The construction is therefore important in discussions of adjoint functors and universal properties.

6.3 Polynomial-like and formal-sum constructions

Free vector spaces resemble polynomial and formal-sum objects because both allow symbolic combinations without imposing extra relations. They are useful when working with formal expressions, combinatorial generating functions, or bases indexed by arbitrary sets. In each case, the emphasis is on algebraic manipulation of symbols rather than on numerical evaluation.

Several algebraic constructions modify the idea of a free vector space by adding relations or extra structure. These variants preserve the general theme of generating a linear object from symbols while changing the rules they must satisfy. Comparing them helps clarify what “free” means in different settings.

7.1 Free vector space with additional structure

Sometimes a free vector space is equipped with extra operations or gradings that reflect additional combinatorial or algebraic information. For example, one may consider graded versions in which basis elements are assigned degrees. The underlying vector space remains free, but the added structure organizes its basis and maps.

7.2 Quotients of free vector spaces

Many familiar vector spaces arise as quotients of free vector spaces by subspaces of relations. In this approach, one begins with formal generators and then identifies combinations that should be equal. This method is common in presentations of vector spaces and in constructions defined by generators and relations.

7.3 Comparison with free algebras

Free algebras generalize the same idea from vector spaces to algebraic systems with multiplication. Whereas a free vector space only supports linear combinations, a free algebra also allows products of generators, subject to the chosen algebraic laws. The free vector space can thus be seen as the linear foundation for more elaborate free constructions.