Basis: Vector Space vs. Topology
In linear algebra and in topology, a basis is a collection of building blocks from which every object of interest can be built. A basis of a vector space is a set of vectors from which every vector is built as a unique linear combination. A basis for a topology is a collection of open sets from which every open set is built as a union (not necessarily unique). A basis of a vector space has no redundant vectors, but a basis for a topology can contain sets that are unions of other sets in the basis.
Definitions
Consider a space of vectors, such as the plane, in which vectors can be added and stretched. A basis of the space is a set of vectors from which every vector in the space can be built in exactly one way, by stretching each of them and adding the results.
Consider a vector space $V$. A basis of $V$ is a set of vectors in $V$ that is linearly independent and spans $V$.
Examples
-
The vectors $(1, 0)$ and $(1, 1)$ in $\reals^2$
A basis need not be perpendicular: $(x, y) = (x - y)(1, 0) + y (1, 1)$.
-
The vectors $(1, 0)$ and $(0, 1)$ in $\reals^2$
The vectors $(1, 0)$ and $(0, 1)$ form a basis of $\reals^2$, because every vector $(x, y)$ is $x (1, 0) + y (0, 1)$, in exactly one way. This basis is the standard basis.
Non-examples
-
The set of all vectors in $\reals^2$
The set of all vectors in $\reals^2$ is not a basis of $\reals^2$, because the set is not linearly independent.
-
Cartesian coordinates $(x, y)$ on a plane, given by an origin $O$ and two perpendicular unit vectors $e_1$ and $e_2$
Cartesian coordinates are not a basis, because they are a rule for labeling points, not a set of vectors. They are built from the basis $(e_1, e_2)$ together with the origin $O$.
-
Latitude and longitude on the surface of the Earth, without the poles
Latitude and longitude are not a basis, because the surface of the Earth is not a vector space, so there are no basis vectors to combine.
-
The vectors $(1, 0)$ and $(2, 0)$ in $\reals^2$
Both point along the $x$-axis, so no combination of them gives $(0, 1)$.
-
Polar coordinates $(r, \theta)$ on the plane without the origin, such that $r > 0$ and $0 \leq \theta < 2 \pi$
Polar coordinates are not a basis, because they are a rule for labeling points, not a set of vectors. They come from no basis, because their coordinates are not linear: doubling a point doubles $r$ but leaves $\theta$ unchanged.
-
The vectors $(1, 0)$, $(0, 1)$, and $(1, 1)$ in $\reals^2$
The vectors $(1, 0)$, $(0, 1)$, and $(1, 1)$ are not a basis of $\reals^2$. They span $\reals^2$, but they are not independent: $(1, 1) = (1, 0) + (0, 1)$, so $(1, 1)$ is redundant.
Properties
- Every vector $v$ in $V$ can be written in exactly one way as a linear combination $v = c_1 b_1 + \dots + c_n b_n$ of the basis vectors. The numbers $c_1, \dots, c_n$ form the coordinate vector of $v$.
- Every vector space has a basis (assuming the axiom of choice), and all bases of the same space have the same number of vectors: its dimension.
- A vector space has many bases. The matrix that represents a linear map depends on which bases are chosen.
Consider a set $X$ and a collection $\mathcal{B}$ of subsets of $X$. The collection $\mathcal{B}$ is a basis on $X$ if
- every point of $X$ is in at least one set in $\mathcal{B}$, and
- for all $B_1, B_2 \in \mathcal{B}$ and every point $x \in B_1 \cap B_2$, there is a set $B_3 \in \mathcal{B}$ such that $x \in B_3 \subseteq B_1 \cap B_2$.
The topology generated by $\mathcal{B}$ is the collection of all unions of sets in $\mathcal{B}$.
Properties
- A collection $\mathcal{B}$ of open sets is a basis for $\mathcal{T}$ if and only if every open set is a union of sets in $\mathcal{B}$.
- A topology has many bases. Every topology $\mathcal{T}$ is a basis for itself, and adding open sets to a basis produces another basis.
- The open balls of a metric space form a basis for its topology.
- A function $f : X \to Y$ between topological spaces is continuous if and only if the preimage under $f$ of every set in a basis for the topology of $Y$ is open in $X$.
- A topological space is called second countable if its topology has a countable basis.
Examples
| Example | basis (of a vector space) | basis (of a topology) | Why |
|---|---|---|---|
| The vectors $(1, 0)$ and $(0, 1)$ in $\reals^2$ | ✓ | – | |
| The set of all vectors in $\reals^2$ | ✗ | – | |
Missing example: open_intervals_in_reals_for_the_standard_topology | |||
Missing example: all_open_sets_of_reals_for_the_standard_topology | |||
Missing example: closed_intervals_in_reals_for_the_standard_topology | |||
Notes
- Both meanings require that every object be built from the basis, but only the vector-space meaning requires that each object be built in exactly one way. Removing a vector from a basis of a vector space produces a set that does not span the space, but removing a set from a basis for a topology can produce another basis.