Dimension as a Test for Isomorphism
Two vector spaces can have very different-looking elements and still have the same linear structure. A vector in \(\mathbb{R}^3\), a polynomial of degree at most \(2\), and a column of three coefficients are different kinds of objects, but each can be specified by three real numbers. The dimension records how many coordinates are needed relative to a basis, and for finite-dimensional real vector spaces it gives a complete answer to whether the spaces are isomorphic.
Throughout this tutorial, spaces and linear maps are over \(\mathbb{R}\). An isomorphism from \(V\) to \(W\) is a linear map \(T:V\to W\) that is both one-to-one and onto. The spaces are isomorphic, written \(V\cong W\), if such a map exists. A basis is an ordered list that is linearly independent and spans the space; the dimension of a finite-dimensional space is the number of vectors in any basis.
This theorem separates the decision from the construction. To rule out an isomorphism, it is enough to show that dimensions differ. To establish that an isomorphism exists, equality of dimensions is enough; matching bases gives a concrete map. The conclusion depends on both spaces being finite-dimensional and over the same scalar field.
Proof. Suppose first that \(T:V\to W\) is an isomorphism. Since \(T\) is one-to-one and onto, the dimension-preservation result for isomorphisms gives \(\dim V=\dim W\). Conversely, suppose \(\dim V=\dim W=n\). Choose bases \(\mathcal{B}=(\mathbf{b}_1,\dots,\mathbf{b}_n)\) for \(V\) and \(\mathcal{C}=(\mathbf{c}_1,\dots,\mathbf{c}_n)\) for \(W\). Each vector in \(V\) has a unique expression \(\sum_{i=1}^{n}a_i\mathbf{b}_i\). Define \(T:V\to W\) by \(T(\sum_{i=1}^{n}a_i\mathbf{b}_i)=\sum_{i=1}^{n}a_i\mathbf{c}_i\). Uniqueness of basis coordinates makes this definition unambiguous. The formula shows that \(T\) is linear. If \(T(\sum a_i\mathbf{b}_i)=\mathbf{0}\), then \(\sum a_i\mathbf{c}_i=\mathbf{0}\); independence of \(\mathcal{C}\) forces every \(a_i=0\), so \(T\) is one-to-one. For any \(\mathbf{w}\in W\), write \(\mathbf{w}=\sum a_i\mathbf{c}_i\); then \(\mathbf{w}=T(\sum a_i\mathbf{b}_i)\), so \(T\) is onto. Thus \(T\) is an isomorphism. \(\square\)
Counting Dimensions Before Comparing Spaces
Dimension comparisons are only useful when the counts are correct. Let \(P_n\) denote the real vector space of polynomials of degree at most \(n\), including the zero polynomial. Its standard basis is \((1,x,\dots,x^n)\): every polynomial in \(P_n\) has a unique expression \(a_0+a_1x+\cdots+a_nx^n\). Thus \(P_n\) has \(n+1\) basis vectors, not \(n\).
The real matrix space \(\mathbb{R}^{m\times n}\) consists of all \(m\)-by-\(n\) matrices. It has \(mn\) entries, and a basis is given by the matrices with a \(1\) in one position and \(0\) elsewhere. There is one basis matrix for each entry position, so its dimension is \(mn\).
For the polynomial count, the powers \(1,x,\dots,x^n\) span because every polynomial of degree at most \(n\) is a linear combination of them. They are independent because a polynomial equal to the zero polynomial has every coefficient equal to zero. For the matrix count, each matrix is the sum of its entries times the corresponding single-entry matrices. Such a sum is the zero matrix only if every entry coefficient is zero. These spanning and independence facts justify the counts rather than relying only on how the spaces are written.
Worked Example: Comparing \(P_2\) and \(\mathbb{R}^3\)
Determine whether \(P_2\) and \(\mathbb{R}^3\) are isomorphic, and describe an isomorphism.
The polynomial space has basis \((1,x,x^2)\), so \(\dim P_2=3\). The standard basis of \(\mathbb{R}^3\) has three vectors, so \(\dim\mathbb{R}^3=3\). The classification theorem therefore guarantees an isomorphism. In coefficient coordinates, define \(T:P_2\to\mathbb{R}^3\) by
For example, \(q(x)=2-3x+4x^2\) maps to \(\begin{bmatrix}2\\-3\\4\end{bmatrix}\). The inverse coordinate assignment sends \(\begin{bmatrix}a\\b\\c\end{bmatrix}\) to \(a+bx+cx^2\). In particular, applying the inverse to the image of \(q\) gives \(2-3x+4x^2=q(x)\), and applying \(T\) to the inverse of \(\begin{bmatrix}2\\-3\\4\end{bmatrix}\) returns \(\begin{bmatrix}2\\-3\\4\end{bmatrix}\). The formulas also show directly that both maps preserve addition and scalar multiplication, so this is an isomorphism.
The example illustrates why the elements’ appearance is not the test. Polynomials are not column vectors by definition, but their unique coefficients provide coordinates. Equal finite dimensions guarantee that the coordinate matching can be made in both directions.
Matrix Spaces and Polynomial Spaces
Worked Example: Comparing \(\mathbb{R}^{2\times2}\) and \(P_3\)
Decide whether the space of real \(2\)-by-\(2\) matrices is isomorphic to the polynomials of degree at most \(3\). Give a coordinate map and verify it on a specific matrix.
There are four entry positions in a \(2\)-by-\(2\) matrix, so \(\dim\mathbb{R}^{2\times2}=2\cdot2=4\). The basis \((1,x,x^2,x^3)\) gives \(\dim P_3=4\). The spaces are isomorphic. One map is
For the matrix \(\mathbf{A}=\begin{bmatrix}1&-2\\3&0\end{bmatrix}\), the map gives \(T(\mathbf{A})=1-2x+3x^2\). The inverse map sends \(a+bx+cx^2+dx^3\) to \(\begin{bmatrix}a&b\\c&d\end{bmatrix}\). Applying it to the polynomial found above gives \(\begin{bmatrix}1&-2\\3&0\end{bmatrix}\), the original matrix. Conversely, applying \(T\) to the inverse of \(a+bx+cx^2+dx^3\) gives \(a+bx+cx^2+dx^3\) again. Both assignments are linear because they carry sums and scalar multiples of coordinates to the corresponding sums and scalar multiples. They are inverses, so the map is a bijection and hence an isomorphism.
Different coordinate conventions could arrange the four entries or coefficients in another order. That would give a different explicit isomorphism, but the dimension comparison would be unchanged. For a yes-or-no decision, constructing the map is optional once the dimensions have been established; when a map is requested, matched bases provide a reliable way to define it.
Unequal Dimensions Rule Out Isomorphism
Worked Example: Comparing \(\mathbb{R}^3\) and \(P_3\)
Determine whether \(\mathbb{R}^3\) and \(P_3\) are isomorphic.
The standard basis gives \(\dim\mathbb{R}^3=3\). But \(P_3\) has basis \((1,x,x^2,x^3)\), so \(\dim P_3=4\). Since \(3\ne4\), the classification theorem implies that no linear bijection between these spaces exists. In particular, matching the first three coordinates of a vector to the first three coefficients of a polynomial cannot give an onto map to \(P_3\): polynomials with an \(x^3\) term, such as \(x^3\), would not be reached.
Check: the obstruction is a dimension difference, not a failure to find a clever formula. The theorem rules out every possible isomorphism, not only the coefficient assignment just considered.
A dimension mismatch does not prevent linear maps from existing; it prevents a map from being both one-to-one and onto. For example, a linear map from a smaller-dimensional space might reach some vectors in a larger one without reaching all of them. Isomorphism requires both properties simultaneously.
Common Mistake: Counting the Wrong Number of Coordinates
A frequent error is to count the degree label instead of the number of basis elements. For instance, one might claim that \(\dim P_3=3\) because the degree bound is \(3\), and then conclude that \(P_3\cong\mathbb{R}^3\). This calculation is wrong: the powers start at \(x^0=1\), so the basis is \((1,x,x^2,x^3)\), which has four members. The correction is \(\dim P_3=3+1=4\), and therefore \(P_3\not\cong\mathbb{R}^3\).
Similarly, a matrix space is counted by the number of independent entry positions, not by its largest matrix dimension alone. A \(2\)-by-\(3\) matrix has six independent entries, so \(\dim\mathbb{R}^{2\times3}=6\), not \(2\) or \(3\). Before using the classification theorem, identify a basis or justify a dimension formula for each space.
| Space | Dimension | Reason |
|---|---|---|
| \(\mathbb{R}^k\) | \(k\) | Its standard basis has \(k\) vectors |
| \(P_n\) | \(n+1\) | Basis \((1,x,\dots,x^n)\) |
| \(\mathbb{R}^{m\times n}\) | \(mn\) | One independent coordinate per entry |
Using the Classification Efficiently
When a question asks whether two spaces are isomorphic, begin by naming the scalar field and confirming that both spaces are finite-dimensional. Then find each dimension, taking care with conventions such as the degree bound in \(P_n\). The classification theorem finishes the decision. If the dimensions match and an explicit correspondence is useful, choose bases and send the \(i\)-th basis vector of one space to the \(i\)-th basis vector of the other. If the dimensions differ, no construction attempt is needed: the dimension-preservation result already rules out an isomorphism.
This criterion is powerful because it ignores the particular nature of the objects while retaining exactly the information relevant to linear structure. Vectors, polynomials, and matrices can be compared through coordinates, and their dimensions tell us whether a linear bijection can exist. The criterion does not classify particular linear maps or determine whether a specified map is bijective; those are separate questions about the map itself.
Check Your Understanding
For each comparison, state the dimensions first and then decide whether the spaces are isomorphic.
- Are \(P_4\) and \(\mathbb{R}^5\) isomorphic? Give the basis dimension that supports your conclusion.
- Find the dimension of \(\mathbb{R}^{2\times3}\). Is it isomorphic to \(P_5\)?
- Explain why \(\mathbb{R}^2\) and \(P_2\) cannot be isomorphic, even if a linear map between them can be defined.
- Describe an isomorphism from \(P_1\) to \(\mathbb{R}^2\) using polynomial coefficients.
- Why does equality of dimensions guarantee the existence of an isomorphism, but not make every linear map between the spaces an isomorphism?
